The accurate assessment of the wind speed distribution plays an important role in wind power estimation. The appropriate probability density function is chosen to reduce the estimation error of wind energy for accurate assessment of wind speed and wind energy in Yunnan mountainous areas, in order to determine most suitable distribution for wind speed data in mountainous areas of Yunnan, a variety of conventional distributions have been proposed for modelling. The maximum likelihood theory is used for parameter estimation. A comprehensive evaluation criterion is proposed, including five statistical determination coefficients R2, KolmogorovSmirnov test, chisquare test, root mean square error (RMSE) and Bayesian information criterion (BIC) for goodnessoffit test, Histograms and QQ plots are used for intuitive analysis, and the relative errors in wind power density are used for comparison. The results show that the Weibull distribution fits the wind speed in the mountainous areas of Yunnan best. This conclusion provides the basis for next accurate assessment of wind energy in mountainous areas of Yunnan.