Construction of scientific and accurate forecasting model for predicting the number of Yunnan schoolage children has certain reference value for the rational planning of the number of primary school, the development of primary school teachers recruitment plan, and the formulation of population policy. According to the principle of minimum sum of prediction error square, a precision of the smoothing coefficient is set up, the optimal smoothing coefficient is selected from all the values between 0 and 1, by using the Matlab loop statement programming. Research shows that the three exponential smoothing model established by this method has good prediction effect on the number of schoolage children. On this basis, the forecasting results of the three exponential smoothing model are modified by using Markov forecasting method, which makes up for the forecast errors caused by random fluctuation. Through the comparison and analysis of the predicted results, it shows that the forecasting effect of cubic exponential smoothingmarkov model is better than cubic exponential smoothing model. Finally, the number of 20142016 Yunnan schoolage children is forecasted by using cubic exponential smoothingmarkov Model.