[Purposes] For the detection data format of LIDAR and camera sensors is not unified, the resolution is different, and the fusion calculation complexity of data level and feature level is high, a target fusion detection method at decision level is proposed in this paper. [Methods] The installation position of LIDAR and camera is calibrated jointly to realize the coordinate system transformation of the detection results of the two kinds of sensors; The Hungarian algorithm is used to match the LIDAR point cloud detection target frame with the camera image detection target frame, set the coincidence area threshold of the target frame, and detect the position and type of the target. [Findings] Through the real vehicle test, the results show that the target recognition accuracy of vehicles and pedestrians is improved by 3.3% and 5.3% respectively. The proposed fusion method is verified by using the public dataset KITTI. The results show that the detection accuracy of the proposed fusion method can reach 75.42%, 69.71% and 63.71% in different difficulty level scenarios， respectively. Compared with the existing common fusion methods, the detection accuracy is improved. [Conclusions] The overlapping area threshold of the detection target frame of the two types of sensors has a great impact on the decision level fusion's detection results. Selecting different thresholds according to the length width ratio of the detection target detection frame can effectively improve the target recognition accuracy of vehicles and pedestrians. The decision level fusion method can accurately match the detection targets of LIDAR and camera, improve the target detection accuracy effectively.