Abstract: Object monitoring is one of the key technologies of logistics information platform system. The study introduced an application method of machine vision technology in the logistics center monitoring system, improved covariance matrix algorithm to monitor objects in the logistics center. Against the technical difficulties of objects detection, the covariance matrix algorithm was applied to monitor objects in the logistics center and against the shortcomings of covariance matrix algorithm in the process of monitoring objects, the study proposed a method of path prediction and template dynamic adjustment. Experiments show that the method can effectively monitor objects in the logistics center, the improved method can not only adapt quickly to pose and scale variations of objects, but also track accurately and continuously those temporarily occluded objects, has good robustness. The method provides a new solution of monitoring objects in the logistics system.