The identification of mycobacterial species in clinical isolates is essential for making patient care decisions. Polymerase chain reaction (PCR) restriction enzyme analysis (PRA) is a simple and rapid identification method, based on amplification of 441 bp of the hsp65 gene and restriction with BstEII and HaeIII. As a contribution to the validation of PRA, a multicenter study was performed in eight laboratories located in Argentina, Brazil, Colombia, Chile, and Guadeloupe. Each laboratory received 18 coded isolates from the collection of the Institute of Tropical Medicine (Antwerp, Belgium), representing duplicates of nine laboratory strains: Mycobacterium terrae CIPT 140320001, Mycobacterium scrofulaceum CIPT 140220031, Mycobacterium flavescens ATCC 14474, Mycobacterium triviale ATCC 23292, Mycobacterium nonchromogenicum ATCC 19530, Mycobacterium chitae ATCC 19627, Mycobacterium abscessus ATCC 19977, Mycobacterium kansasii ATCC 12478, and Mycobacterium peregrinum ATCC 14467. A detailed protocol including amplification, enzymatic digestion, and gel preparation was provided to each laboratory. Two laboratories identified correctly all 18 (100%) isolates, one identified correctly 17 (94.5%), two identified 14 (77.7%), one identified 11 (61%), and two identified 8 (44.4%) isolates. Errors detected in laboratories with more than 77% accuracy were associated with electrophoresis running conditions and an unspecific amplicon produced by a single strain. Lower accuracy was mainly related to inappropriate use of DNA markers and insufficient training in interpretation of patterns. In conclusion, the PRA method was readily implemented in some Latin American and Caribbean laboratories of mycobacteria, but improvements in critical points, as gel running conditions and training in interpretiation of patterns, are needed in order to improve accuracy. In others, improvement in critical points is still necessary.