A deep knowledge of tire behavior in operating conditions is fundamental to the effective modelling of vehicular dynamics on its safety, comfort and performance aspects. Data-based models are a common approach despite the associated challenges: the complex interaction between tire constructive and operational factors implies the necessity of large datasets, the tradeoff between local and global fit is challenging and the handling of a high number of inputs with varying relevance to each output is a computationally expensive problem.A very promising approach to data-based modelling is the Gaussian Process Regression (GPR), a class of supervised learning. Data points are used to train an underlying probability distribution with characteristics assumed a priori. The resulting model has relatively small requirement of training data, robustness against overfitting, good response to complex behavior and computational tractability.The aim of this work is to support the elaboration of data-based tire models by creating one of a Formula SAE specific 10" slick tire. Procedures are presented for the use of GPR to fit the data locally and then predictions are made on lateral and longitudinal forces with respect to vertical load, slip-angle, slip-ratio, pressure and camber. In the end, model quality metrics will be established for internal cross validation and comparison to test data.