Prediction using different machine learning approaches have been applied in last few decades in different areas. Waiting time is an undeniable fact for every queue and it is very important to develop a system that predicts its duration in real life with minimum error. In this paper we applied several machine learning algorithms and among them we chose Support Vector Regression (SVR) in a real life Banking queue dataset that contains real-life queues of multiple Banks where we predicted waiting time for each individual in the queue. Moreover, we have compared the result of prediction using SVR with different classifications and clustering methods such as K-nearest-neighbor and K means Clustering. We have shown the feasibility of applying SVR in prediction of waiting time in banking queues of developing countries for each individual, which is applicable and it performs well in queue analysis.