The study of ocean waves is a very difficult task for engineers and scientists. The complexity of the actual phenomena is substantial. The mathematics involved, especially in the more elaborate models, are similarly complex. The available deterministic mathematical models differ considerably from the chaotic reality of ocean waves. In the 40’s and early 50’s statistics were getting involved in the study of ocean waves and specifically in the study of the distribution of wave height and period. A major breakthrough was achieved in 1955, when a linear stochastic approach was introduced by Pierson (Pierson and Marks, 152). This model, is probably the main approach in the study of ocean waves today. An important characteristic of the theory is the distribution of energy over the frequency-direction space, characterized by a function called the directional spectrum. The estimation of the directional spectrum I very important, especially for engineering design, and several methods have been developed for this task. The project presented here is a small contribution to this research, relative to a new improvement in the Maximum Entropy method, which uses a recently developed technique in optimization problems called Simulated Annealing.