Woodrats of the genus Neotoma are an important study system for ecological and paleoecological research. However, paleontological studies are often hindered by the difficulty of identifying woodrat remains to species. We address this limitation by using 2D landmark-based geometric morphometrics to classify 199 lower first molars (m1s) of five extant western North American Neotoma species (N. albigula, N. cinerea, N. fuscipes, N. lepida, and N. macrotis) collected throughout California. We then use discriminant analysis of principal components (DAPC) models to identify Late Pleistocene fossils of unknown species from the Rancho La Brea Tar Pits in Los Angeles, California. DAPC correctly identifies ∼85–90% of extant individuals to species, with most misclassifications occurring between sister taxa N. fuscipes and N. macrotis. Most fossil m1s are classified as N. macrotis by DAPC, which may be the first confirmation of N. macrotis in the fossil record. We show that landmark-based geometric morphometric analyses are generally effective at differentiating m1s of extant Neotoma species in California and they are an auspicious method for unknown fossil identification. Further applications of this method across a broader range of geographic locations and species will better contextualize its utility.