The invention discloses a method for predicting yield and quality of the alfalfa grass, and relates to the technical field of agriculture. The method of the invention comprises: building the regression model by acquiring the growth time of the alfalfa grass and the indicator data of yield and nutritional quality corresponding to the growth time; or, building the regression model by acquiring the greater or equal to 0C accumulated temperature of the alfalfa grass and the indicator data of yield and nutritional quality corresponding to the accumulated temperature; predicting the yield and nutritional quality of the alfalfa grass in the next year and determining the best harvest time of the alfalfa grass by putting the growth time or the greater or equal to 0CC accumulated temperature of the alfalfa grass to be analyzed into the corresponding regression model. The method of the invention is characterized by less climate effect, simpleness, practicability, rapidness, convenience, accuracy and low cost; therefore, this method is suitable for agricultural product enterprises, agricultural cooperatives, individual farmers and other small and medium-sized production units. This method can save a lot of money and time for small and medium-sized agricultural production units. To a certain extent, the production cost of forage can be reduced, and the production efficiency of forage can be improved.