Abstract: Aiming at the problem of high detection rate and false detection rate of abnormal data in complex time-varying topological networks, an optimization method for extracting abnormal data from complex time-varying topological networks is proposed. The density space clustering method is introduced to calculate the local anomaly factors of data sets according to the density reachability relationship between data sets. On the basis of acquiring local anomaly factors, the number of data points in the detection area is simplified, the complexity of the calculation process is reduced, and the neighborhood optimization of anomaly data detection in time-varying topological networks is realized. Combining with information gain, the feature attributes of data are distinguished, and the feature selection optimization objective function is constructed to optimize the feature attributes of anomaly data detection. The experimental results show that the performance of the improved anomaly data detection algorithm is significantly improved, which guarantees the high detection rate and low false alarm rate of the algorithm, and the running time is shorter.