OBJECTIVE To evaluate and compare the accuracy of root canal lengths (RCLs) in primary teeth using different clinical methods. METHODS RCLs of 70 extracted human primary single-rooted teeth were estimated in vivo by a calibrated investigator using the electronic apex locator (EAL) method, radiovisiography (RVG), conventional radiography (X-ray), and tactile sensation method (TSM). The teeth were categorized into Group 1 (G1; teeth without physiological root resorption [PRR]); Group 2 (G2; teeth with one-fourth apical PRR); and Group 3 (G3; teeth with one-fourth to three-fourths lingual PRR). RCLs determined by various in vivo methods were compared with the ex vivo actual root canal length method (ARCL), as determined by stereo-microscope. Data obtained were compared using analysis of variance, Newman-Keuls test, and Student's test. RESULTS In Groups 1 and 2, the mean RCL determined by EAL was closest to that of the gold standard ARCL, followed by RVG, X-ray, and TSM. In G3, the mean RCL determined by EAL was closest to ARCL, followed by TSM, RVG, and X-ray. CONCLUSIONS A fifth-generation electronic apex locator was found to be useful in accurately determining root canal lengths of primary teeth with or without physiological root resorption.