Information fusion technology is a multidisciplinary highly integrated popular research fields. In view of the security threats in the coal mine environmental monitoring system, A distributed hierarchical information fusion algorithm is proposed in wireless sensor network. In the source, the inproved Kalman filtering are used to reduce the power consumption and data collision. In the sink, the dynamic weighted fusion algorithm based on linear minimum variance sense is adopted. This method is minimize the overall power consumption and the probability of collision in the network as well, it can realize signal reconstruction. A simple example is performed to verify the reliability and fusioning precision of this method, at the same time it has high fault-tolerant.