Abstract: The handling behaviour of skid-steered tracked vehicles is more complex than conventional wheeled vehicles because of their non-linear characteristics. One of the traditional methods used for establishing the handling of wheeled vehicles is the constant radius test, where the variation of steer angle with lateral acceleration is investigated. There have been no comparable studies either experimentally or theoretically for a tracked vehicle for this type of test. With the introduction of a variable-steer-ratio (ratio of sprocket angular speeds) system on tracked vehicles it may be possible to perform this test experimentally. This study addresses the problem of producing a theoretical prediction for the constant-radius test for a current in-service tracked vehicle, Combat Vehicle Reconnaissance (Tracked) CVR(T) manufactured by Alvis in the UK. A model of a general skid-steer tracked vehicle is developed and validated against experimental data from a trial on CVR(T). The model is then used to predict the results for a constant radius turn for this vehicle. The results show that the CVR(T) initially understeers before going into oversteer.