In motor sports, telemetering is often used to evaluate a racing car's condition by monitoring various characteristics. Especially in an endurance race, accurate judgments of conditions are required because an error by technical personnel can lead to retirement from a race. It is, therefore, considered that constructing a vehicle abnormality detection system is effective for enhancing the accuracy of judgments made by technical personnel and competitiveness. In constructing the system to detect abnormalities in a vehicle during a race, we investigated the possibility of applying MT system for data obtained in the 1999 Le Mans 24-Hour Race and Le Mans Fuji 1000 km Race. As a result, we were able to verify the utility of the vehicle abnormality detection system using data from actual races, and show that application of the MT system in the motor sports field would be useful for improving competitiveness. For the analysis, which is also useful in other fields, it is effective to divide a unit space by characteristic values, and it is possible to incrementally increase the unit spaces in accordance with race progress as the number of laps increases. In response to the problem that dispersion becomes zero with in a unit space, we set the allowable fluctuation range from characteristic values and measuring accuracy, and changed the data within that range, thus making it possible to use all of the characteristic values.