The invention discloses a large-scale power abnormal data detection method and system based on machine learning. The invention discloses a large-scale power abnormal data detection method. The methodcomprises the following steps: 1) inputting a user data record to be detected; 2) extracting related features of the user data; 3) inputting the features into a multi-layer perceptron neural network model; and 4) outputting the detection probability, and adjusting the parameters of the neural network model of the multi-layer perceptron based on the minimized loss function. The multi-layer perceptron neural network model capable of detecting the abnormal power data is adopted, elements with a causal relationship with the abnormal data acquisition are mined on the basis of massive acquired datacontaining abnormal information, the abnormal state of the acquired power data is explored, and the general type of data exception can be effectively detected.