In view of problem of high power consumption and data generation rate of existing mine-used accurate positioning system, a cognitive algorithm for mine-used accurate positioning system was designed. The recognition algorithm reasonably reconfigures transmitter power of target node and location data sampling interval by sensing changes of moving speed and position of target node and using reconstruction method of decision-making, achieves optimization of system power and data generation rate. The simulation results show that the cognitive algorithm can effectively reduce power consumption and data generation rate of accurate positioning system without affecting track performance of positioning system to node trajectory.