Objective: To compare the retentions of different designs of cobalt-chromium (Co-Cr), pure titanium (CP Ti), and titanium alloy (Ti-6Al-4V) removable partial denture (RPD) circumferential clasps manufactured by selective laser melting (SLM) and to analyze the stress distribution of these clasps during the removal from abutment teeth. Methods: Clasps with clasp arm size A (1.9 mm width/1.1 mm thickness at the body and 0.8-taper) or B (1.2 times A) and 0.25 mm or 0.50 mm undercut engagement were modeled on a prepared first premolar die, named as designs A1, A2, A3, and A4, respectively. The density and elastic modulus of SLM-built Co-Cr, CP Ti, and Ti-6Al-4V were measured and given to different groups of clasps. The density, elastic modulus, and Poisson’s ratio of enamel were given to the die. The control group was the cast Co-Cr clasp with design A1, to which the density and elastic modulus of cast Co-Cr alloy were given. The Poisson’s ratio of all metals was 0.33. The initial 5 N dislodging force was applied, and the maximum displacement of the clasp along the insertion path was computed. The load was reapplied with an increment of 5 N than in the last simulation until the clasp was completely dislodged. The retentive force range of different groups of clasps was obtained. The retentive forces of the SLM-built Co-Cr, CP Ti, and Ti-6Al-4V clasps with equivalent computed retentive force range to the control group were validated through the insertion/removal experiment. The von Mises stress distributions of these three groups of SLM-built clasps under 15 N loads were analyzed. Results: SLM-built Co-Cr, CP Ti, and Ti-6Al-4V clasps with designs B1 or B2, and Co-Cr clasps with design A2 had higher retentive forces than those of the control group. SLM-built CP Ti and Ti-6Al-4V clasps with design A1 had lower retentive forces than those of the control group. SLM-built Co-Cr clasp with design A1 and SLM-built CP Ti and Ti-6Al-4V clasps with design A2 had equivalent retentive forces to those of the control group. The insertion/removal experiment showed that the measured retentive forces of these three groups of SLM-built clasps were (21.57±5.41) N, (19.75±4.47) N, and (19.32±2.04) N, respectively. No statistically significant measured retentive force difference was found among these three groups of SLM-built clasps (P>0.05). The maximum von Mises stress of these three groups of SLM-built clasps exceeded their responding yield strength except for the Ti-6Al-4V one. Conclusion: SLM-built Co-Cr circumferential clasps had higher retention than CP Ti and Ti-6Al-4V ones with the same clasp arm size and undercut engagement. The retention of SLM-built circumferential clasps could be adjusted by changing the undercut engagement and clasp arm size. If SLM-built circumferential clasps are used in clinical practice, the Ti-6Al-4V clasp with clasp arm size A and 0.50 mm undercut engagement is recommended considering the long-term use of RPD in the patient’s mouth.