Aiming at fifteen indices of the population the gross regional production, the fiscal revenue and expenditure, agriculture, industry, service, the annual precipitation of Hunan province from the years of 2000 to 2010, as well as the historical data on the area of Hunan’s public welfare forest, this paper firstly decided the economy & population development level and the natural environment level as the two main ingredients of the variation of public welfare forest with the method of main ingredient analysis, then both the quantitative forecast of the total demand of Hunan’s public welfare forest through analyzing the scores of above two main ingredients based on the multiple linear regression model and its rectified quantitative forecast based on the cubic curve nonlinear regression model were carried out. The results showed the effectiveness of this model, which could carry out rolling rectifying of the demand of public welfare forest and provide scientific reasons for the control and forecast of the demand of public welfare forest.