In the article the features of energy summation from two wind generators, located at a certain distance from each other, are considered. The method of calculating the correlation function between the wind flow speed change functions in the direction of wind distribution is presented. The formulas for describing the fluctuation components of energy at the output of the wind generator are given for two cases: when the phases of the fluctuations of the wind flow on two wind generators are the same and when the fluctuations of the wind flow are in the antiphases. It is shown that to increase the energy level that can be taken from the wind power plant it is necessary to control the phase shift between the energy fluctuations at the output of the wind generators and use the energy of the storages; and to use linear approximations to approximate the wind speed change function. Under the condition of a linear change of the internal resistance of the wind generator in time, it is advisable to introduce the wind speed change function with linear approximations. The system of orthonormal linear functions based on Walsh functions is given. A table with formulas and graphs describing the first 8 functions, which are arranged in order of increasing the number of their sign alternating on the interval of functions definition, is presented. The result of the approximation of the wind speed change function with a system of 8 linear functions based on Walsh functions is shown. Decomposition coefficients, mean-square and average relative approximation errors for such approximation are calculated. In order to find the parameters of multiple linear regression the method of least squares is applied. The regression equation in matrix form is given. An example of application of linear regression prediction method to simple functions is shown. The restoration result for wind speed change function is shown. Decomposition coefficients, mean-square and average relative approximation errors for restoration of wind speed change function with linear regression method are calculated.