It is said that a new method (Maximum Entropy Method or MEM) for power spectral density estimation is superior to other conventional methods such as Blackman and Tukey or FFT. This technique is especially valuable for the short data samples and has much resolution. Making use of such merit, we have calculated the wind wave data especially in the currents. Consequently, it is possible to calculate the power spectral density, but gives less resolution for the short data. In order to improve the resolution it is required to increase the number of the prediction error filter. But, a large number of it gives the unstable spectrum. It means that the degrees of freedom of the spectral estimate is decreasing. It we have enough data length, such problem became out of the question. In this case the peak frequency and the intergrated power are estimated with great accuracy.