When a scene contains a similar background to the object, the traditional tracking method based on sparse representation may produce the drifting problems, a new object tracking method is proposed. Based on the local binary pattern features(LBP) of the object, the object appearance model can sparse representation simultaneously with the original object template and the template built by some particles in the current frame, and build a joint objective function which can solve the tracking issue through an iterative optimization problem. Experimental results show that the track method can well track with occlusion and illumination issues, as well as the scene sequence with a similar background to the object.