A prospective clinical study to compare for the first time the accuracy of digital and conventional maxillary implant impressions for completely edentulous patients.Sixteen patients received maxillary implant supported fixed complete dentures. After the verification of the conventional final casts, the casts were scanned with a desktop (extraoral) scanner. Intraoral full-arch digital scans were also obtained with scan bodies and STL files. Extraoral and intraoral scans were superimposed and analyzed with reverse engineering software. The primary outcome measure was the assessment of accuracy between scans of the verified conventional casts and digital full-arch impressions. The secondary outcome was the effect of the implant number on the 3D accuracy of impressions with Spearman's rank correlation coefficient.The 3D deviations between virtual casts from intraoral full-arch digital scans and digitized final stone casts generated from conventional implant impressions were found to be 162 ± 77 μm. In the 4-implant group, 5-implant group, and 6-implant group the 3D deviations were found to be 139 ± 56 μm, 146 ± 90 μm, and 185 ± 81 μm, respectively. There was a positive correlation between increased implant number and 3D-deviations, but there was no statistically significant difference (p = 0.191).The 3D accuracy of full-arch digital implant scans lies within previously reported clinically acceptable threshold. Full-arch digital scans and a complete digital workflow in the fabrication of maxillary fixed complete dentures may be clinically feasible.