The method of short-term risk assessment for distribution network with microgrids is studied. System states are selected using the improved state sampling method, which simplifies the sampling process and takes the sequential characteristics into account. Models of loads and DG output used in this paper can describe their variability conveniently. Method of minimum time segmentation is applied to accurately calculate the state consequences. On this basis, power and economic risk indicators are used to evaluate the risk of system. The RBTS BUS6 system is used as an example for 24 h short-term risk assessment. Results demonstrate that configuration of microgrids contributes to a reduction on load-loss and economic risk of distribution network.