The invention discloses a new energy plant station network-related end real-time interaction process anomaly detection method and system, and the method comprises the steps: carrying out the analysisof a message data packet of a new energy plant station network-related end application layer based on a preset feature vector, and obtaining each feature vector data corresponding to the message packet; based on eigenvector data, comparing a K-NN algorithm with the sample data of various types of abnormal messages to obtain the class of the message data packet. According to the scheme, based on the preset feature vectors, the feature vector data in the message data packet is purposefully acquired for common category analysis, so that the analysis efficiency is improved. Feature fields of the message data are compared with the sample data of abnormal data and the categories of the message data is compared and determined by the K-NN matching algorithm The judgment method is simple and easy to implement, parameters do not need to be estimated, training is not needed, the method is suitable for rare events and multi-classification problems, matching of the multi-classification characteristics of the new energy plant station is effectively achieved, and the safety protection level of a new energy plant station system is effectively improved.