The invention discloses a method for calculating remanufactured part environmental loss based on a back propagation (BP) neural network. The method comprises the following steps of: (1) acquiring a measured part, wherein values of remanufacturing process parameters which influence the measured part serve as detection data; (2) establishing an error BP neural network model according to the detection data which are normalized; (3) training and testing the error BP neural network model, if an error is smaller than an error threshold, executing the step (4), if the error is not smaller than the error threshold, continuously training and testing the error BP neural network model, and finishing process until the number of iterations reaches preset number and the error is not smaller than the error threshold; (4) checking a network by using a second part of data, judging whether the error BP neural network model meets a requirement on matching rate, if so, executing the step (5), and otherwise, executing the step (3); and (5) inputting actual detection data of a remanufactured part into the error BP neural network model which meets the requirement on the matching rate, and predicting the environmental loss.