Upper body power (UBP) is an important determinant of cross-country ski race performance. Although numerous studies exist to measure UBP of cross-country skiers, to date, no study has ever attempted to predict UBP of cross-country skiers. The purpose of this paper was to develop prediction models for estimating 10-s UBP and 60-s UBP of cross-country skiers using support vector machines (SVM). Four types of SVMs have been considered, they are as follows: SVM using the radial basis function kernel (SVM-RBF), SVM using the sigmoid kernel, SVM using the polynomial kernel, and SVM using the linear kernel. For comparison purposes, UBP prediction models based on multilayer perceptron and multiple linear regression have also been developed. The dataset used in this study includes data of 77 subjects. Age, gender, height, weight, body mass index, maximal heart rate, maximal oxygen uptake, and exercise time are the predictor variables, and and are the target variables. Several UBP prediction models have been developed by using the combination of the predictor variables to predict and . By using 10-fold cross-validation on the datasets, the performance of the models has been evaluated by calculating their standard error of estimates (SEEs) and multiple correlation coefficients (Rs). The results show that SVM-RBF-based UBP prediction models perform much better (i.e., yield lower SEEs and higher Rs) than the prediction models developed by other regression methods and can be safely used for the prediction of UBP of cross-country skiers.