Image target recognition is one of the most important topics in the fields of image processing and pattern recognition. It has been more and more widely used in the fields of security monitoring, medical diagnostics and so on. Traditional algorithm can accurately identify the non-occluded target with simple background. For the partially-occluded target, however, the traditional recognition algorithms often make mistakes of false alarm and missed alarm. To solve the problems caused by partial occlusion, this paper proposes a modified target recognition algorithm, based on matching feature points. This algorithm gets the feature points roughly by corner detection of Harris, estimates the effectiveness of the feather points' probability density, therefore effectively recognize the target. The proposed algorithm has better real-time tracking, and it provides a good solution to solve the recognition of the partially-occluded target.