Uncertainty in projections of global change impacts on biodiversity over the 21st century is high. Improved predictive accuracy is needed, highlighting the importance of using different types of models when predicting species range shifts. However, this is still rarely done. Our approach integrates the outputs of a spatially-explicit physiologically inspired model of extinction and correlative species distribution models to assess climate-change induced range shifts of three European reptile species (Lacerta lepida, Iberolacerta monticola, and Hemidactylus turcicus) in the coming decades. We integrated the two types of models by mapping and quantifying agreement and disagreement between their projections. We analyzed the relationships between climate change and projected range shifts. Agreement between model projections varied greatly between species and depended on whether or not they consider dispersal ability. Under our approach, the reliability of predictions is greatest where the predictions of these different types of models converge, and in this way uncertainty is reduced; sites where this convergence occurs are characterized by both current high temperatures and significant future temperature increase, suggesting they may become hotspots of local extinctions. Moreover, this approach can be readily implemented with other types of models.