In order to solve the problems of low average accuracy of power consumption data identification and poor accuracy of anti-intermittent electricity stealing behavior monitoring, an anti-intermittent electricity stealing behavior monitoring method based on neural network algorithm is proposed. Constructing a basic electricity-stealing analysis model, using a median filter to eliminate acquired useless data, completing the acquisition and preprocessing of power consumption data, applying a neural network back propagation algorithm to optimize an electricity-stealing behavior monitoring neural network model, setting an intermittent electricity-stealing behavior identification function, and realizing the monitoring of the intermittent electricity-stealing behavior; The experimental link was constructed, and the application effect of this method was analyzed by F1 value and average accuracy. The experimental results show that this method improves the ability of data analysis, and further improves the accuracy of anti-intermittent electricity theft monitoring.