Abstract: A time series wind speed model is presented based on continuous state markov chain (CSMC) in this paper. The recorded wind speed time series is transformed into that in the normal distribution domain using Box Cox technique, and then a CSMC model of transformed data is constructed in multi dimensional normal domain. In addition, normal distributed data is transformed into wind speed data by using of the Box Cox back transformation, therefore a CSMC model of wind speed time series is built. The validity of proposed model is testified by measured wind speed data, which shows that the model can preserve the probability distribution and autocorrelation characteristics of original wind speed. The applicability of the proposed model is also verified on the IEEE RTS test system, which shows that it can be used to evaluate the reliability of power systems with wind energy.