Objective: To provide a reference for using intraoral scanners for making clinical diagnostic dentures of edentulous jaws by comparing the accuracy of three intraoral scanners for primary impression and jaw relation record of edentulous jaws. Methods: This study contained 6 primary impressions of the edentulous patients. Each of the impressions consisted of the maxillary primary impression, the mandibular primary impression and the jaw relation record. For each of them, a dental cast scanner (Dentscan Y500) was used to obtain stereolithography(STL)data as reference scan, and then three intraoral scanners including i500, Trios 3 and CEREC Primescan were used for three times to obtain STL data as experiment groups. In Geomagic Studio 2013 software, trueness was obtained by comparing experiment groups with the reference scan, and the precision was obtained from intragroup comparisons. Registered maxillary data of the intraoral scan with reference scan, the morphological error of jaw relation record was obtained by comparing jaw relation record of the intraoral scan with the reference scan. Registered mandibular data with jaw relation record of intraoral scan and the displacement of the jaw position were evaluated. Independent samples t test and Mann-Whitney U test in the SPSS 20.0 statistical software were used to statistically analyze the trueness, precision and morphological error of jaw relation record of three intraoral scanners. The Bland-Altman diagram was used to evaluate the consistency of the jaw relationship measured by the three intraoral scanners. Results: The trueness of i500, Trios 3 and CEREC Primescan scanners was (182.34±101.21) μm, (145.21±71.73) μm, and (78.34±34.79) μm for maxilla; (106.42±21.63) μm, and 95.08 (63.08) μm, (78.45±42.77) μm for mandible. There was no significant difference in trueness of the three scanners when scanning the maxilla and mandible(P>0.05). The precision of the three scanners was 147.65 (156.30) μm, (147.54±83.33) μm, and 40.30 (32.80) μm for maxilla; (90.96±30.77) μm, (53.73±23.56) μm, and 37.60 (93.93) μm for mandible. The precision of CEREC Primescan scanner was significantly better than that of the other two scanners for maxilla(P<0.05). Trios 3 and CEREC Primescan scanners were significantly better than i500 scanner for mandible(P<0.05). The precision of the i500 and Trios 3 scanners for mandible was superior to maxilla(P<0.05). The upper limit of 95% confidence intervals of trueness and precision of three scanners for both maxilla and mandible were within ±300 μm which was clinically accepted. The morphological error of jaw relation record of the three scanners was (337.68±128.54) μm, (342.89±195.41) μm, and (168.62±88.35) μm. The 95% confidence intervals of i500 and Trios 3 scanners were over 300 μm. CEREC Primescan scanner was significantly superior to i500 scanner(P<0.05).The displacement of the jaw position of the three scanners was (0.83±0.56) mm, (0.80±0.45) mm, and (0.91±0.75) mm for vertical dimension; (0.79±0.58) mm, (0.62±0.18) mm, and (0.53±0.53) mm for anterior and posterior directions; (0.95±0.59) mm, (0.69±0.45) mm, and (0.60±0.22) mm for left and right directions. The displacement of the jaw position of the three scanners in vertical dimension, anterior and posterior directions and the left and right directions were within the 95% consistency limit. Conclusion: Three intraoral scanners showed good trueness and precision. The i500 and Trios 3 scanners had more errors in jaw relation record, but they were used as primary jaw relation record. It is suggested that three intraoral scanners can be used for obtaining digital data to make diagnostic dentures and individual trays, reducing possible deforming or crack when sending impressions from clinic to laboratory.