Abstract: This paper considers the situation where multi-mobile sinks and path endpoints are located along the edge of the circumference, and abstracts it as a hybrid optimization problem characterized in high dimensionality and large searching space. Classic algorithms like the k-splitour algorithm cannot optimize its continuous variables. This paper first obtains k sub-paths by adopting k-splitour algorithm and designs the method to eliminate the crossing of sub-paths to acquire local optimum for discrete variables. Then the algorithm acquires multi-mobile sinks path planning results efficiently by designing local optimization methods for continuous variables to decide on the location of access points on each communication disk. The upper bound of the algorithm and its theoretical proof are presented. The experiments show the effectiveness of both the designed model and its algorithm in solving the path planning problem in data collection.