Pipe bursts occur in water supply networks, and the identification quick detection of bursts is a long-standing problem for water supply companies. For the rapid detection of burst pipes in metering sub-district water supply networks of district metering area, a new prediction-classification-correction three-stage Infomer-Zscore algorithm is proposed. The Infomer-Zscore algorithm solves the problem of low data processing efficiency and abnormally low water consumption in traditional methods. In the prediction stage, using the deep learning Informer algorithm to predict long sequence time-series Long Sequence Time-series of pressure data for the pipe network, improving the accuracy of water pressure prediction and the efficiency of data processing. In the classification stage, improved robustness of water pressure data over time by using a multi-threshold classification method .The Infomer-Zscore algorithm achieved a 90.9% true positive rate (TPR) ,a 1.7% false positive rate (FPR) and a 99.5% detection accuracy (DA) in the pipe burst simulation test. Long sequence Sequence timeTime-series pressure forecasting can be used not only for burst identificationdetection, but also for effective pressure control in the network to reduce the risk of bursts.