In view of problem that flight path and trajectory of four-rotor flying robot in coal mine roadway are changeable, which leads to great redundancy of environment information data collected by the robot in dynamic space, an environmental information data compression algorithm for coal mine four-rotor flying robot is proposed. The default threshold noise reduction method of wavelet analysis is used to reduce the noise of original data to improve signal-to-noise ratio; taking methane monitoring as an example to introduce data compression algorithm, the roadway space is split into several sections along the roadway direction, and the effective value of methane concentration of each split section is selected and used to reconstruct the detection data of methane concentration. The experimental results show that the methane concentration signal compressed and reconstructed by the algorithm has little error with the manual detection result, the data compression algorithm can extract the effective value of methane concentration data and remove the redundant environmental information collected by the robot, thereby improving the effectiveness and real-time performance of data transmission.