To solve the problem of human localization and tracking in the service space of human-robot coexistence, a novel method based on collaborative perception using distributed laser radars for human tracking is proposed. First, laser data obtained from each of radars is evaluated statistically, separately. All data are divided into static data and dynamic data. Static data are employed to calibrate the laser radars for background subtraction. Dynamic data are then sent to server computer by wireless network. Data from different radars at the same moment are formed into one frame to realize data synchronization and integration. Finally, contour model matching based interative closest point algorithm is implemented on each frame of laser data to distinguish the targets. To realize human tracking, targets between two frames are associated using association gate based on their positions and speeds. Experimental results demonstrate the effectiveness of the proposed method in solving the problem of human localization and tracking. Compared to vision-based tracking methods, the system performs better in terms of localization accuracy and tracking success rate.