Abstract: In order to providing a reliable screening method for port supervision departments to quickly identify the quality of unknown imported coal，the construction research of imported coal quality evaluation model was carried out based on BP neural network.The model was based on SPSS principal com⁃ ponent analysis，finding that 8 parameters including low calorific value of received base，high calorific val⁃ ue of received base，high calorific value of dry base，total water，ash of dry base，mercury of dry base，ar⁃ senic of dry base and phosphorus of dry base，can replace the original 15 parameters to evaluate the quali⁃ ty of imported coal. By using the above parameters，59 groups of imported coal samples were randomly se⁃ lected as the training set，and the other 13 groups of samples as the verification set，and the BP neural net⁃ work model of imported coal quality was constructed. The predicted result shows that relative errors be⁃ tween the predicted value of the model and the actual value by test are less than 15%，meaning that the quality of unknown imported coal could be rapidly identified according to the predicted value of the BP neural network model.

Key words: imported coal, principal component analysis, neural network model, quality evaluation