We previously proposed an accurate cup placement procedure using a portable navigation system (PNS) for total hip arthroplasty (THA) in the lateral decubitus position; however, the accuracy of our proposed procedure has not been shown, the aim of this study was to demonstrate the accuracy. We prospectively analyzed 79 hips treated with primary THA; 40 hips treated until June 2020 were included in the conventional procedure (CP) group, and 39 hips treated from July 2020 were included in the modified procedure (MP) group. In the MP, pelvic orientation is considered to be the set coordinate axes in addition to the CP using the PNS. The accuracy was based on the difference between the navigation record (NR) and postoperative computed tomography measurement. The radiographic inclination (RI) and anteversion (RA) accuracies were 1.55° and 2.14°, respectively, in the MP group and 3.03° and 6.20°, respectively, in the CP group (p < 0.001). The error was within 5° of the NR for both the RI and RA in 34 in the MP group (87.2%) and 14 in the CP group (35.0%) (p < 0.001). The error was within 5° of the target angle (RI 40°, RA 15°) for both the RI and RA in 29 hips in the MP group (74.7%) and 12 in the CP group (30.0%) (p < 0.001). Our procedure with the consideration of pelvic orientation achieved dramatically improved the accuracy of PNS and was suitable to facilitate accurate cup placement.