Aiming at the multiple uncertain features of wind speed with both randomness and ambiguity, a distributed wind power bilevel programming model for active distribution network (ADN) based on stochastic fuzzy chance constraint is proposed. Considering the multiple uncertainty characteristics which contains both randomness and fuzziness of wind speed, firstly, the article describes the uncertainty of distributed wind power output through random fuzzy simulation, established a random fuzzy model of wind power output by considering the time sequence,obtains the chance measure of static security index by random fuzzy simulation and power flow analysis, uses reliability which contains both randomness and fuzziness to evaluate the distributed wind power configuration scheme. The model describes the uncertainty of distributed wind power output through random fuzzy simulation, and considers its timing characteristics to establish a random fuzzy model of wind power output. Random fuzzy simulation and power flow analysis are combined to obtain system static security such as node voltage, branch power and system reverse power of random fuzzy chance measure. The credibility of both randomness and fuzziness is adopted as the index to evaluate the distributed wind power capacity allocation plan. Secondly, the system power balance, static safety index opportunity measurement constraints and active management (AM) measures are considered. The maximization of annual profit random fuzzy expected value is set to be the upperlevel optimization goal. The minimization of random fuzzy expected value of the distributed wind power is set to be the lowerlevel optimization goal and realized by the active management of distributed wind power. Whether the static safety index of the upperlevel scheme meets the confidence level of the random fuzzy chance constraint, the upper and lower levels are planned to collaborate to construct a new bilevel programming model for distributed wind power capacity configuration with random fuzzy chance constraint. Finally, a combination of random fuzzy simulation, forward and backward power flow calculation and genetic algorithm (GA) is proposed to solve the model. The simulation process and results of IEEE14 node example show the effectiveness and superiority of the model and algorithm.