Purpose: The purpose of this investigation was to evaluate the effects of number and distribution of implants upon in vitro dislodging forces to a simulated implant-supported overdenture and to examine differences between several different attachment systems. Materials and Methods: An experiment was undertaken utilizing a model simulating a mandibular edentulous ridge with dental implants in positions on the model approximating tooth positions in the natural dentition. A cobalt-chromium-cast testing framework was used to measure the peak load required to disconnect an attachment. Four different types of commercially available attachments were used in various positions on the model in sequence to evaluate the effects of retention and stability of overdentures based on implant number and distribution: (1) ERA, (2) O-Ring, (3) Locator, and (4) Ball. For each group, 10 measurements were made of peak dislodging forces. Means were calculated and differences among the systems, directions, and groups were identified using a repeated measured analysis of variance (α = .05). Results: The interactions between the attachment system, direction of force, and implant number and distribution were statistically significant. Vertical dislodging forces of the simulated overdenture prosthesis increased with additional widely spaced implants. Oblique dislodging forces of the simulated prosthesis increased with additional widely spaced implants except in the two-implant model with all attachments, and in the four-implant groups with Locator attachments. Anteroposterior dislodging forces of a simulated overdenture prosthesis increased with additional widely spaced implants except in the four-implant groups with Ball and Locator attachments. Ball attachments reported the highest levels of retention and stability followed by Locator, O-Ring, and ERA. Conclusions: Within the limitations of this study, retention and stability of an implant overdenture prosthesis are significantly affected by implant number, implant distribution, and abutment type.