Aiming at the trajectory fusion problem of multi-target multi-platform tracking, a trajectory fusion algorithm based on feature matching is proposed. Compared with the single camera for target tracking, the fusion of multiple camera information can provide richer trajectory information, make up for the blind area of the single camera, and improve the robustness of target detection and tracking. The paper solves the problem of how to obtain a more accurate multi-target fusion trajectory under multi-source information. The paper performs matching and alignment of multi-target observation trajectories under multiple platforms through feature matching, and derives the same target multi-track fusion method based on error minimization. The multi-platform multi-target tracking simulation experiment is carried out by using the vehicle tracking in the urban environment. Using the method proposed in this paper, the multi-platform observation trajectory of the same target is fused. Data description of simulation experiment shows that compared with the single-platform observation trajectory, the multi-platform observation trajectory fusion results not only have higher positioning accuracy, but also achieve cross-platform tracking.