This paper establishes early-warning system for NR market risk based on the fact that BP artificial neural network is the warning model and the producing region’s price of NR is the alarming index. This paper also uses the annual rate from 1995 to 2008 to train BP network, employs weight matrix and transfer function to calculate the sensitivity of all index and analyzes its impact from economics theory, which indicates that the theory is in conformity with results and the choices of index are reasonable.The result shows that the established warning system can identify the risk of NR market and reflect the uncertainty by economic crisis and important policy changes.