Abstract: Rare earth is an important strategic resource in China.Understanding the price fluctuation of rare earth is very important for efficient utilization of resources.The BP neural network model based on principal component analysis is adopted on the basis of the price factors of rare earth resources.The prices of rare earth products are predicted according to the combined model.There are many factors that affect the price of rare earth products.In this paper, principal component analysis is used to eliminate the redundant information among the influencing factors.The dimension of BP neural network input data is reduced to improve the prediction accuracy.This paper takes dysprosium oxide price as the forecasting object, chooses the monthly data from January 2010 to February 2018, and constructs a multi-factor PCA-BP combination model.The prediction results show that the combined model is superior to the traditional BP neural network model in simulation capability, error level and convergence speed.Combined models can predict dysprosium oxide prices more accurately.