Academic performance during the first year of university is correlated with future academic success, and is considered to be a determining factor in the reduction of dropouts. In the present study, we describe a new academic performance indicator for the first term of first‐year science degrees students at La Laguna University in Spain. We are interested in identifying the most important previous academic factors for predicting the success of first‐year students at university. Specifically, multiple linear regression models were used to identify such predictors of academic success. We report that, for all of the analyzed science degrees, the best predictor of academic success is high school grade point average. In addition, we obtained predictive models for estimating the value of the new academic performance indicator. Using these models, we can predict future academic success, which may help identify students at risk of failure at the beginning of the course. This in turn would ensure early implementation of educational interventions or strategies to increase academic achievement of such students.