The single feature tracking algorithm of imaging guidance system has a disadvantage of feature instability which can cause the shift of tracking point and the loss of targets in the course of tracking. A multi-feature joint matching method for target tracking was proposed. The drawbacks of tracking algorithm using gray or point feature were analyzed, the corresponding relationship between target tracking algorithm and feature type was induced , and several multi-feature modes were given according to the experiences. The validity of the method using gray and point features as multi-feature was demonstrated. The algorithm is use the gray feature and point feature of target image to calculate the tracking points, respectively. The consistency between the tracking point of each tracking algorithm and the ideal target points was determined according to the measure criteria of comparability between tracking point neighbor and template,a weight coefficient was allocated for each matching algorithm, and an optimal tracking point was given. Besides, the tracking point was taken as a basis of template update. The algorithm overcomes the shortcoming of single tracking method and realizes the stabilized target tracking under the complex background. At last, the tracking video data of a certain missile was used to demonstrate the validity of multi-feature tracking algorithm, compared with a single feature method, the algorithm combines the virtues of multiple features, and the calculated tracking point shift is small. And also, the further improvement of tracking trajectory smoothness and the real-time implementation of algorithm were investigated.