This paper mainly describes studies hydrophone placement strategy in a complex underwater environment model to compute a set of "good" locations where data sampling will be most effective. Throughout this paper it is assumed that a 3-D underwater topographic map of a workspace is given as input.Since the negative gradient direction is the fastest descent direction, we fit a complex underwater terrain to a differentiable function and find the minimum value of the function to determine the low-lying area of the underwater terrain.The hydrophone placement strategy relies on gradient direction algorithm that solves a problem of maximize underwater coverage: Find the maximize coverage set of hydrophone inside a 3-D workspace. After finding the maximize underwater coverage set, to better take into account the optimal solution to the problem of data sampling, the finite VC-dimension algorithm computes a set of hydrophone that satisfies hydroacoustic signal energy loss constraints. We use the principle of the maximize splitting subset of the coverage set and the ”dual” set of the coverage covering set, so as to find the hitting set, and finally find the suboptimal set (i.e., the sensor suboptimal coverage set).Compared with the random deployment algorithm, although the computed set of hydrophone is not guaranteed to have minimum size, the algorithm does compute with high network coverage quality.