Abstract: It was a difficult problem in infrared target tracking field that how to effectively exclude the interference of infrared decoy, especially for that infrared decoy technique had developed more and more advanced as a common measure to interfer infrared tracking capability. Based on discussing characteristic difference of infrared targets and decoys, a multi-target tracking strategy to track excluding the infrared decoy was firstly proposed. And an improved joint probabilistic data association(JPDA) algorithm with fusion of multiple features was established. Finally, the process of tracking with excluding infrared decoy was simulated. Simulation results showed that the proposed algoritllIn can track target at real-time, which has higher accuracy for resolving the problem of the infrared target recognition that infrared decoys form false targets and the target block.