Synthetic wind speed time series are widely used in the planning and assessment of the wind power integrated system.The original first-order Markov chain approach is not able to retain the autocorrelation property of observed wind speed data.Meanwhile，it is impossible for it to reflect the variation of wind speed due to changes of season and weather.An advanced first-order Markov chain approach is then put forward，in which the seasonal characteristics and diurnal behavior of wind are modeled by dividing a year into 12 and a day into4equal parts，respectively.Also，the relationship between wind and precipitation is taken into account.A comparison between the advanced approach and the original one shows that this advanced first-order Markov chain approach can not only retain the autocorrelation property of observed wind speed data fairly well，but it has improved the accuracy of the wind series generated as well.