The deployment of Wireless Sensor Networks to support optimal coverage is obviously a fastidious task, especially in hard to reach or inaccessible area. To address this problem, simplifying assumptions and simulations are commonly used techniques. However, these techniques may not be able to capture all the details about the problem to be solved. Hence, transition from the simulation to the actual deployment causes a loss of accuracy and robustness in coverage problem. In this paper, we address the above issue using two main ideas. First, pre-processing step with terrain analysis in three-dimensional environment. Second, a robust modeling approach based on both average and variance of visibility from each cell. We formulate the coverage optimization problem as a decision problem, where the objective is to maximize the area covered by a fixed number of sensors. Numerical results show that the proposed techniques are more efficient than the standard-based model, especially in hard to reach areas.