The Key Largo cotton mouse (Peromyscus gossypinus allapaticola) was federally listed as Endangered in 1984 due to habitat loss and subsequent population decline, but because of infrequent monitoring, the extent of decline and long-term population trends are unknown. We modeled population abundance under a variety of trapping scenarios to develop a monitoring protocol that would provide accurate population estimates and could be conducted more frequently under current personnel and budget constraints. We captured cotton mice on 33 trapping grids in Key Largo, Florida, USA, during 3 trapping sessions (1 Mar–11 May, 4 Jul–7 Sep, and 29 Oct–31 Dec) in 2007. We compared demographic parameter estimates from subsets of trapping grids with estimates from all grids to identify a subset of grids that would produce population estimates within 10% and remain within the 95% confidence intervals of estimates using all grids when both methods were extrapolated to total available habitat. There were insufficient captures in session 1 to produce a reliable population estimate. We found that 13 and 12 trapping grids, respectively, produced abundance estimates meeting our criteria in sessions 2 and 3. We conclude that the subset of 12 grids trapped during November and December will allow a 64% reduction in number of trapping grids necessary to provide a reliable population estimate. Employed annually, our recommended trapping protocol provides state and federal agencies an efficient method for monitoring trends in the Key Largo cotton mouse population. © 2013 The Wildlife Society.