The high complexity of shape, attitude and size of the infrared images of the aerial aircrafts may result in low recognition rate and processing speed to the classical infrared target recognition methods. To solve the problem, we proposed a method for the aircraft target recognition based on infrared images and feature fusion. The proposed method made full use of the algorithms of scale invariant feature transform and singular value decomposition. The recognition features of infrared image were extracted at first, and then the fusion feature vector was constructed. The aircrafts were recognized by BP neural network. Based on infrared image database of five types of aircraft, we made experiments for comparision of our method with the classifical methods. The result show the effectiveness and feasibility of the method.