Tyres are essential components of a race car, being the only direct link between the vehicle and the road. In racing environments, it is crucial to model them with a high accuracy level to maximize the performance of the whole vehicle. Common tyre models, such as Magic Formula, are described by a pure mechanical behaviour, neglecting the influence of parameters such as temperature, which has a big effect on force production capability of tyres. Because of that this master thesis proposes an adaptive thermal tyre model, aimed to improve the estimation of tyre forces produced in the contact patch. First, a tyre surface temperature estimator is proposed, which allows to measure this parameter without sensors and a low computational cost.