Purpose: Clinical scientific evidence of the accuracy of intraoral scanning (IOS) impressions for multiple implant-supported prostheses in partially edentulous arches is limited. This in vivo study aimed to evaluate the accuracy of IOS impressions for multiple implants by comparing them with analog impressions. Materials and Methods: Patients with three adjacent implants in partially edentulous arches were recruited. For each arch, IOS impressions were obtained. Using the analog splinted impression technique, casts were obtained and scanned with an extraoral scanner. Using metrology software, the distances and angles between the implants were measured in all IOS and analog impression Standard Tessellation Language files. The absolute values of the distance and angular deviation were calculated and analyzed by comparing the two impression methods. Results: Thirty-four patients with partially edentulous arches were enrolled. The distance deviation between the IOS and analog impressions was 85.09 ± 58.73 µm, which was significantly lower than the clinically acceptable error (150 µm; P < .001). The angular deviation was 0.515 ± 0.426°. Distance deviations were significantly correlated with the interimplant distances (r = 0.384, P = .001) and angulations (r = 0.278, P =.022). Conclusions: For multiple implant-supported prostheses in partially edentulous arches, the distance deviation between the IOS and analog impression was within the clinicall acceptable range of misfit and increased with increasing interimplant distance and angulation.