The sea surface targets are densely located and have diverse types, and the large amount of clutter false alarm points may cause high error association rate in tracking processing. In order to solve this problem, a sea surface targets association algorithm based on speed partition and multiple information utilization is proposed. First, the targets are divided into three categories according to their speed, respectively corresponding to different speed ranges. In each speed range, the corresponding range gate, speed gate and association rules are set up, where different association rules are adopted based on the feature information of targets, such as shape, size, amplitude and compact degree, etc. Then, the sea surface target track is filtered. The processing results of collected data show that the algorithm can obtain good tracking effect.