Optimistic after a string of laboratory successes, researchers are assembling potential vaccines against AIDS and may start preliminary testing within a year, scientists at the Sixth International Conference on AIDS said today.

In sessions so crowded that some people were turned away, researchers on Friday reviewed their most hopeful work, steps they have taken toward creating a vaccine that could prevent spread of the human immunodeficiency virus.

''In the past year, we've cracked open the door in our optimism about a vaccine,'' said [Dr. Wayne C. Koff (born 1952)], chief of vaccine research at the National Institute of Allergies and Infectious Diseases. ''And I think in the next two or three years we will knock it down.''

Some of the experiments came after dozens of failed experiments in animals since 1986, failures that produced a deep pessimism about whether a vaccine would ever be possible.

Pregnant Women as Subjects

Reports given at the meeting demonstrated that monkeys, chimpanzees, cats and cells in an artificial human immune system can be protected against infection. These early experiments will continue, Dr. Koff said at a news conference after the Friday session, but it is also urgent to begin tests as soon as possible.

He said that within a year, researchers may begin experiments in which a vaccine will be used to treat pregnant women already infected with the virus. The researchers hope the vaccine may retard the development of disease in the women and may prevent them from passing the infection to their babies, Dr. Koff said. About one-third of the babies born to infected women acquire the disease in the womb.

Tests of the more conventional use of a vaccine, to protect the population against infection, are not likely to begin for two to four years. Researchers said how long it would take to create a working vaccine depended on success at each step along the way, though there is now hope that one may be on the market in 10 years.

Progress and Difficulties

Vaccine research was the topic of several sessions at the conference over three days ending today. [Dr. Jay Arthur Berzofsky (born 1946)], chief of immunogenetics and vaccine research at the National Cancer Institute, outlined the progress and difficulties to those at the conference Friday and elaborated on those points in an interview today.

The recent successes demonstrate that a vaccine in humans is at least feasible, Dr. Berzofsky said.

Results from three successful chimpanzee experiments and three successful monkey experiments were reported at the meeting, but Dr. Berzofsky said the success was limited. In each experiment, the animals were given a vaccine, then given doses of the live virus, and the researchers found that the animals were protected against infection by the virus.

Two Types of Infection

Dr. Berzofsky said that the doses of live virus given to the animals were low and that the virus was given in shots, whereas actual human infection occurs through the skin of the genitals or elsewhere.

Also, he said, only the virus itself was given to the animals, while humans are infected by such free virus as well as virus hidden inside infected cells in the blood or sperm. The body has different means of responding to each type of infection.

It is not known if the animal tests would be successful if they were given both free virus and cell-carried virus, Dr. Berzofsky said.

[Dr. Jay A. Levy (born 1938)], a virologist at the University of California at San Francisco, said, ''That will be one of the most important experiments to do now. When animals are protected from cell-associated virus, that will be a real breakthrough,'' he said.

Successes and Failures

One series of experiments reported here explored the difference between the failed experiments of previous years and the successful ones of the past year.

Researchers from Genentech, a California biotechnology company, reported that in 1986 they had taken a portion of the outer shell of the virus and injected it to make the chimpanzees' immune system produce antibodies and killer cells as if a viral attack was under way.

But after the vaccination, when live virus was given to the chimp, the protection failed and the animal quickly became infected.

Beginning in 1989, the group changed its approach. Researchers around the country began to report that a crucial segment of the viral shell, called the Third Variable region, or V3 loop, is the most powerful in creating defenses against the virus.

The Genentech group found that in its earlier experiments that region had been snipped in half by enzyme, and was disabled. In the new experiments they made sure the V3 loop was intact. The experiments succeeded, and two chimpanzees were protected by the new vaccine.

Infection Is Quickened

Researchers at a competing company, [MicroGeneSys, Incorporated] of Boston, say there are other important active sites. In addition, each different strain of the virus has slightly different versions of these active sites, so that it will probably be necessary to assemble a vaccine that includes bits of active sites from five or six different strains.

The Genentech team also noted a potential danger in the vaccine work in general: When the immune system makes antibodies to protect itself, some antibodies, ''enhancing antibodies,'' actually do the opposite and may dramatically speed up the progession of disease.

In a chimpanzee experiment that failed to protect the animals, Genentech researchers saw that the animals became infected twice as fast as might normally be expected.

In order to overcome the problem, the researchers say, a vaccine must not include bits of the shell that will elicit ''enhancing antibodies.''

About 30 Vaccine Candidates

There are now about 30 vaccine candidates being tested in laboratories around the world, and about $180 million is being expended on the work, Dr. Koff said, including $150 million for research and $30 million for preliminary safety tests being done in humans.

At the meeting, two groups, one led by Dr. Jonas Salk, creator of the Salk polio vaccine, and Dr. Alexandra Levine of the University of Southern California, and the other led by Dr. Robert Redfield of the Walter Reed Army Medical Center, reported that giving vaccines to humans produces a substantial immune response without significant side effects.

In the work presented by Dr. Levine, 82 patients already infected with the virus were given a vaccine created from a live but disabled AIDS virus. Among the 82, given the vaccine at various times over the past two years, only one has developed an infection characteristic of AIDS, and five have developed Kaposi's sarcoma, the cancer associated with AIDS.

The experiment had no comparison group, so no conclusions may be drawn, but it appears that rate of progression to disease is about a fifth of the rate found in other experimental groups.