The paper uses daily historical data on electricity prices in the Latvian segment of Nord Pool power market in combination with wind speed measurements conducted in the north-western part of the country in order to perform revenue forecast for a wind power generator park and access the economic feasibility of such projects in Latvia. The forecast of revenues is based on two stochastic differential equation models calibrated on the historical data of wind speed and electricity prices. Pre-testing procedures suggest that mean reverting Vasicek model can be used for wind speed modelling, while electricity price development is better described by a mean reverting model with jump diffusion. Monte Carlo simulation approach is used to generate the distribution of forecasted revenues and estimate project returns taking into consideration the cost structure and technical characteristics of an existing wind power plant in Latvia. The results of the study show that, given the historical market prices of electricity, wind energy potential on the north-western part of Latvia is insufficient to justify the construction of wind energy farms in the region for the considered wind generator type. In the absence of state support, a prudent approach to the selection of wind generator types should be chosen.