In sensor networks with a path-fixed mobile sink, due to the limited communication time of the mobile sink and random deployment of the sensor nodes, it is quite difficult to increase the amount of data collected and reduce energy consumption simultaneously. To address this problem, this paper proposes a data collection scheme called maximum amount shortest path (MASP) to optimize the mapping between members and sub-sinks. MASP is formulated as an integer linear programming problem which is solved by a genetic algorithm. A communication protocol is designed to implement MASP, which is also applicable in sensor networks with low density and multiple sinks. Simulations under OMNET++ shows that MASP outperforms shortest path tree (SPT) and static sink methods in terms of energy utilization efficiency.