The invention discloses a method for predicting the prices of traditional Chinese medicinal materials based on large data calculation. An X12 seasonal adjustment algorithm is adopted to decompose historical monthly prices into trend cycle factors, seasonal factors, irregular factors and seasonal trading day mixed factors of monthly prices. A time sequence method is adopted to predict the obviously regular trend cycle factors to obtain new trend cycle factors, then the seasonal factors of the months are recovered, and the CPI values of the months in a year are represented by the CPI of the current year. Finally, the rainfall data in each year is excavated, and the trend cycle factors, the seasonal factors, the irregular factors, the seasonal trading day mixed factors, the CPI in the current month and the yearly rainfall are used as input layers of an LM-BP neural network for analysis and prediction. By adopting the method disclosed by the invention, the prediction effect is obviously better than that of a single time sequence method, the prediction precision of the monthly price of bighead atractylodes rhizome is improved to a larger extent by the new method, so that the method can be prompted to the market to serve as a method for predicting the monthly prices of the traditional Chinese medicinal materials.