Artificial neural network (ANN) has the characteristics of wide information distribution, high fault tolerance and self-organizing ability. The error back-propagation algorithm (BP) can approximate any continuous function, has a strong nonlinear mapping ability, it is suitable for incomplete information or inference rules in uncertain environment. Mineral resources assets audit responsibility evaluation included some fuzzy feature data, in the evaluation process requires expert knowledge and

more »

... erience, and BP neural network evaluation method can eliminate the influence of human subjective weight setting, which makes the evaluation results more scientific and objective. This paper constructs the mineral resources assets responsibility audit index system which includes 13 indexes, and designs the BP neural network evaluation model with 13 input nodes, 9 hidden layer nodes, and 1 output layer nodes .Through the network training, the results show that the correct recognition rate of the BP neural network improved algorithm can reach 75%.The research is helpful to the exploration of the resource audit evaluation method.