This paper focuses on the problem of seabed survey mission with multiple target points distributed in large scale area. To complete the survey with AUVs in minimum cost, a two-step procedure using greed strategy is presented. In the first step, the locations of so-called anchor points are found by executing k-means algorithm iteratively. In the second step, an ant-cycle system is used to find out the optimized order of access from each anchor point to target points nearby. Simulation is also implemented to prove the validity of the algorithm.