Abstract: Considering the challenges brought by the spatial distribution of drones to the deployment and design of urban hotspot area networks， a heterogeneous network framework for space drones is constructed based on stochastic geometry theory. Firstly， the mixed interference model is built by three-dimensional deployment of air small base stations and two-dimensional deployment of ground macro base stations， and the load balance of each base station is achieved by using the maximum bias received power association strategy； In addition， considering the communication characteristics of drones， the air to ground channel is assumed as Nakagami-m fading model； Finally， the mathematical expression of network outage probability is obtained through the distribution of association probability and communication distance between users and base stations. The simulation results show that the scheme of three-dimensional dense deployment of drones can more realistically reflect the interruption performance of drones in hot service scenarios. At the same time， the network performance can be effectively improved by properly setting the bias factor and increasing the density of drones.