This paper presents a novel multiple targets tracking algorithm which has been tested for passive sonar application in simulated real time surveillance environment. The algorithm caters to all requirements of a tracker and allows user input and automatic tracking for track initiation and track maintenance. The algorithm has multiple target tracking capability, accurate data association, enhanced flexibility and adaptability to the environment and target maneuver rates which are required in helping the sonar operator and command for effective decision making and appropriate allocation of Under Sea Warfare (USW) sources. The prediction techniques used here for data association are also presented. It is a soft tracking algorithm with respect to noise rejection and is adaptive to the targets maneuvers. Performance of the tracker for slow, medium and extreme maneuvers with constantly, gradually and rapidly changing rates has been tested and is considered suitable for passive sonar applications. The tracker is robust and has been tested for data association of multiple targets in cross configuration and cluttered environment. The performance of the tracker has been made adaptive to the continuously changing signal to noise ratio under a preset range. It is made sure that noise does not interfere with tracking quality. The range of signal to noise ratio has been so defined that the stealthier targets are not missed due to their changing aspects and false alarms. The algorithm is simulated in Matlab and implemented in C++.