This paper develops a hybrid measure-correlate-predict (MCP) strategy to assess long-term wind resource variations at a farm site. The hybrid MCP method uses recorded data from multiple reference stations to estimate long-term wind conditions at a target wind plant site with greater accuracy than is possible with data from a single reference station. The weight of each reference station in the hybrid strategy is determined by the (i) distance and (ii) elevation differences between the target farm site and each reference station. In this case, the wind data is divided into sectors according to the wind direction, and the MCP strategy is implemented for each wind direction sector separately. The applicability of the proposed hybrid strategy is investigated using five MCP methods: (i) the linear regression; (ii) the variance ratio; (iii) the Weibull scale; (iv) the artificial neural networks; and (v) the support vector regression. To implement the hybrid MCP methodology, we use hourly averaged wind data recorded at five stations in the state of Minnesota between 07-01-1996 and 06-30-2004. Three sets of performance metrics are used to evaluate the hybrid MCP method. The first set of metrics analyze the statistical performance, including the mean wind speed, wind speed variance, root mean square error, and mean absolute error. The second set of metrics evaluate the distribution of long-term wind speed; to this end, the Weibull distribution and the Multivariate and Multimodal Wind Distribution models are adopted. The third set of metrics analyze the energy production of a wind farm. The best hybrid MCP strategy from 256 different combinations of MCP algorithms and reference stations is investigated and selected. The results illustrate that the many-to-one correlation in such a hybrid approach can provide a more reliable prediction of long-term on-site wind variations than that provided by the one-to-one correlations. The accuracy of the hybrid MCP method is found to be highly sensitive to the combination of individual MCP algorithms and reference stations used. It is also observed that the best combination of MCP algorithms is influenced by the length of the concurrent short-term correlation period. (C) 2014 Elsevier Ltd. All rights reserved.