Differential diagnosis of Entamoeba histolytica (pathogenic) and Entamoeba dispar (nonpathogenic), which are two morphologically identical species of amebae, is essential both for treatment decision and public health knowledge. The study reported here was designed to choose a reference differentiation technique. Stool samples (n = 95) were tested by microscopy, TechLab enzyme-linked immunosorbent assays (ELISAs), and an in-house PCR. The target for the PCR amplification was a small region (135 bp) of the SSU rRNA selected to increase the sensitivity of the test. Sixty-eight specimens tested positive by PCR: 2 for E. histolytica and 66 for E. dispar. For detection of E. dispar, ELISA performance was lower than that of microscopy in this reference context, while PCR was much more sensitive than microscopy. Given the low proportion of E. histolytica cases, test performance for this species is difficult to assess. However, for differentiation, PCR performed well on simulated samples, while ELISA gave a discordant result for one of the two samples PCR positive for E. histolytica during the study. This report also confirms that E. dispar infection is significantly higher among travelers and underlines the possibility of acquiring E. histolytica infection in regions that are not areas of endemicity. Because of its lower sensitivity, the interest of ELISA for Entamoeba detection and differentiation in stools seems questionable in nontropical regions. On the other hand, results suggest that PCR should be useful as a reference test for sensitive differentiation of both species and to contribute to physicians' decision in treatment of E. histolytica- or E. dispar-infected patients.