Abstract: When the voltage of parallel point of DC microgrid falls in a short time due to network side faults and other factors, the grid connected converter station usually output minimum reactive power to keep the active power adjustment margin required for the safe operation of DC microgrid as far as possible However, a large number of cases show that this strategy pays too much attention to the influence of active power margin of converter station on the stability of DC microgrid, and ignores the positive significance of more active and abundant reactive power support on the stability of AC power grid. In order to obtain a more reasonable active-reactive power distribution mode of converter station and corresponding micro-grid control strategy, this paper first constructs a quantitative evaluation model of active power safety margin and reactive power support efficiency of converter station during the voltage sag on the network side. Then, the dynamic constraints in the active power regulation and reactive power support of the DC microgrid are obtained by analyzing the conversion relationship between the active power balance and the DC voltage fluctuation at the system level and the short-term operating constraints of the components in the microgrid at the component level On this basis, The reactive power support efficiency and active power safety margin of converter station were optimized based on INSGA2-DS algorithm. In the process of online application, considering the long calculation time of INSGA2-DS algorithm, it is difficult to adapt to the time scale of voltage sag, a timing optimization strategy of "quasi-synchronous calculation, real-time matching" is proposed. A Pareto curve mapping rule is designed to correct the leading edge curve deviation because of the negative effect of the fluctuation of distributed power supply on the timeliness of quasi-synchronous calculation. Finally, the effectiveness of the proposed model and strategy is verified by a large number of numerical examples.