The workshop on Deep Learning for Generic Underwater Video Tracking (UWGOT 2022) is an exciting chance for the visual object tracking community. UWGOT is a platform for inter-communication and discussion through many presentations of advanced research from the worldwide research communities of computer vision theory, applications, and deep learning. The workshop will be hosted in the conjunction with the Asian Conference on Computer Vision (ACCV-2022). High-quality papers are solicited from the tracking community to promote the tracking benchmarking but are not limited to tracking only. The winners will report their methods and results during the workshop. In this challenge, we propose large-scale Underwater Generic Object Tracking (UWGOT) dataset which consists of 200 video sequences. UWGOT is the second large-scale dataset after UOT100 and consists of quite diverse categories of objects. The underwater sequences are collected from YouTube and a marine laboratory here in the UAE. All 200 video sequences are annotated manually using tightly coupled bounding boxes on each object. Annotations are performed using the computer vision annotation tool. The dataset is divided into training and testing sequences. 70% of the sequences are used for training and 30% of sequences are used in testing. The proposed UWGOT dataset will be made available to the public for fare comparison of SOTA trackers.