Species distribution models (SDMs) can be used to predict the probability of species' occurrence. We tested the use of validated SDMs to estimate the number of occurrences of rare plant species across Southern Ontario. We built SDMs for six rare species using known occurrence records and then surveyed 282 new sites and used presence/absence records from these sites to predict probability of occurrence. We summed these probabilities to estimate the number of occurrences on the landscape. We then used simulation exercises which showed that the true number of extant occurrences can be overestimated with fewer than 1,000 SDM-directed survey sites. Therefore, our estimates may be overestimates of the true number of extant occurrences, and more surveys will be required to obtain more accurate estimates. This technique for estimating the number of rare species occurrences will inform managers as they prioritize time and money towards decisions around species recovery and protection.