Crawler mechanisms have the advantage of stable navigation on uneven terrain; as a result, such mechanisms have been adopted for many types of locomotion of outdoor robots, including "search and rescue robots". However, crawler mechanisms always slip when tracking curved paths, and it generates a large accumulating positioning error in vehicles as opposed to conventional wheeled mobile robots. To measure the velocity of the vehicle correctly and improve the accuracy of the odometry, consideration of crawlers' slippage is very important. In this research, we propose a more accurate odometry method for crawler vehicles. In the proposed method, the vehicle can estimate the slip ratios using information from encoders (attached to the actuators) and gyro-sensors. The validity of the method was confirmed by experiments using our crawler vehicle.