Aiming at the problems of node voltage fluctuation and active power loss increase in active distribution network caused by the intermittence of distributed generation during the optimal scheduling of source-network-load- storage, a low-carbon optimal scheduling method of source-network-load-storage based on improved wind-solar scene clustering combined with virtual energy storage is proposed. First, in order to effectively deal with the challenge of uncertain wind and solar output, an iterative self-organizing data analysis techniques algorithm (ISODATA) based on density-based spatial clustering of applications with noise (DBSCAN) is proposed. The good and bad clustering effects of wind-solar scene are evaluated by DB index, Dunn index and contour coefficient. Secondly, the stepped carbon trading mechanism is introduced, and an optimal scheduling model with the goal of minimizing the comprehensive operation cost of the active distribution network is established. At the same time, two operation evaluation indexes of user participation satisfaction and power grid dependence are proposed. Finally, the example simulation is analyzed in the IEEE 33-node system. The results verify the rationality of the DBSCAN-ISODATA algorithm. The proposed low-carbon optimal scheduling method can effectively reduce the carbon emission and operation cost of the system, and realize the low-carbon economy and stable operation of the active distribution network.