Abstract: This paper focused on the problem of the source localization of coal mine gas release. Due to the existence of wind leakage,it is difficult to find the location accurately by using the conventional algorithms. In this study,a multi- source sensor fusion based algorithm for locating the coal mine gas release was proposed. Firstly,the gas distribution in the goaf of a fully mechanized working face was analyzed and the observation model of mine goaf area and the diffusion model of gas release were established. Then,the gas release source parameters of goaf were estimated using the Mixed Kalman Particle Filter algorithm (MKPF),and the coordinate of the estimated parameters was obtained according to the iterative operation. Finally,the data of wireless sensor target source sensing nodes and cluster head nodes were fused to locate the accurate coordinate of the gas release source. The research concludes that the MKPF algorithm has obvious advantages in accuracy localization,which can effectively overcome the difficulty of gas release source localiza- tion impacted by wind leakage. It can provide support for gas warning and gas drainage in goaf.