Managers require accurate models to predict post-fire tree mortality to plan prescribed fire treatments and examine their effectiveness. Here we assess the performance of a common post-fire tree mortality model with an independent dataset of 11 tree species from 13 National Park Service units in the western USA. Overall model discrimination was generally strong, but performance varied considerably among species and sites. The model tended to have higher sensitivity (proportion of correctly classified dead trees) and lower specificity (proportion of correctly classified live trees) for many species, indicating an overestimation of mortality. Variation in model accuracy (percentage of live and dead trees correctly classified) among species was not related to sample size or percentage observed mortality. However, we observed a positive relationship between specificity and a species-specific bark thickness multiplier, indicating that overestimation was more common in thin-barked species. Accuracy was also quite low for thinner bark classes (<1 cm) for many species, leading to poorer model performance. Our results indicate that a common post-fire mortality model generally performs well across a range of species and sites; however, some thin-barked species and size classes would benefit from further refinement to improve model specificity.