A reasonably-complete bird inventory is the crucial starting point for the analysis of the bird community. We evaluated the efficiency of point counts in detecting forest birds and verified how many sampling points or occasions are needed to adequately characterize the bird community. We sampled birds in 5 forest stands (conifer and beech forests) from northern to southern Italy in 2012. Sampling (through aural and visual clues) lasted 5 minutes, during which species were recorded. Data were analysed in relation to both the number of sampling points and the number of sampling occasions. Then, estimates of species richness were compared to random resampling of subsets of the original data. Results showed that after 3.8 sampling occasions (out of 19–24 sampling points) or 10.4 sampling points (given points are sampled 5 times), the species coverage of each community approached, or exceeded, the 90% threshold. Also, no difference in the mean values emerged with the subset estimates, but the latter appeared less precise. Our results suggest that the density of 1 sampling point per every 5 ha, each repeated at least 3 times, can represent an adequate optimization of the sampling effort. We provided useful methodological information for planning bird inventories in forest environments (applicable at least for Mediterranean and south-European mountain forests) when personnel and financial resources are limited, leading to a thoughtful fund management whilst providing a method to evaluate the reliability of species coverage for bird surveys.