This study presents a framework to assess the wind resource of a wind turbine using uncertainty analysis. Firstly, probability models are proposed for the natural variability of wind resources that include air density, mean wind velocity and associated Weibull parameters, surface roughness exponent, and error for prediction of long-term wind velocity based on the Measure-Correlate-Predict method. An empirical probability model for a power performance curve is also demonstrated. Secondly, a Monte-Carlo based numerical simulation procedure which utilizes the probability models is presented. From the numerical simulation. it is found that the present method can effectively evaluate the expected annual energy production for different averaging periods and confidence intervals. The uncertainty, which is 11% corresponding to the normalized average energy production in the present example, can be calculated by specifically considering the characteristics of the individual sources in terms of probability parameters. (C) 2009 Elsevier Ltd. All rights reserved.