The invention relates to a medium voltage distribution network disconnection ungrounded fault detection method based on machine learning. The medium voltage distribution network disconnection ungrounded fault detection method based on machine learning includes the following steps: S1, extracting daily electricity consumption data from a distribution transformer level; S2, pre-processing the extracted data, including removing of trend items and periodic processing on a time sequence; and S3, classifying the pre-processed data by a random forest algorithm based on feature selection to obtain a fault detection result. Compared with the prior art, based on the distribution and utilization power information system data and the machine learning algorithm, the medium voltage distribution networkdisconnection ungrounded fault detection method based on machine learning uses a pure data-driven method to propose a method for intelligently detecting line disconnection of the distribution network.Through multiple comparisons, the medium voltage distribution network disconnection ungrounded fault detection method is high in the classifying accuracy, and can select the parameters of the main standard for judging disconnection of the distribution network.