A method based on cluster analysis techniques is proposed in order to obtain representative local wind patterns. Modeling renewable energy systems, in an accurate way which strongly depends on wind loads, requires reliable data to analyze their response and evaluate their performance. Otherwise poor agreement with the models may be obtained. In a case study for La Paz, Baja California Sur, Mexico, four different wind patterns were obtained from a one-year data set recorded at 10-min intervals.