Objective To evaluate the prognostic value of surgical treatment in gallbladder squamous cell carcinoma (GSCC) by using real-world data with a large sample in the Surveillance, Epidemiology and End Results (SEER) database. Methods The clinical data of patients with pathologically diagnosed GSCC from 2000 to 2019 were extracted from the SEER database. According to the inclusion and exclusion criteria, a total of 257 patients were included after strict screening. The patients were divided into operation group and non-operation group according to whether they underwent surgery. The cancer-specific survival (CSS) and the overall survival (OS) between the two groups were compared, and the influencing factors for the CSS and the OS were analyzed by using Cox proportional hazard model. Results Of 257 patients, 127 (49.4%) were in the operation group, and 130 (50.6%) in the non-operation group. The average follow-up ranged from 0 to 220 months, with the median follow-up time of 3 months. Of the 127 patients in the operation group, 105 died (82.7%), including 88 tumor-related deaths (69.3%). Of the 130 patients in the non-operation group, 124 died (95.4%), including 115 tumor-related deaths (88.5%). The median survival time for OS in the operation group and the non-operation group were 6 months and 3 months, respectively, and that for CSS were 7 months and 3 months, respectively. The estimated 1-year OS of the operation group and the non-operation group were 30.1% and 4.6% respectively; the estimated 1-year CSS were 35.1% and 5.8%, respectively. There were significant differences between the two groups on OS and CSS (χ2=41.400, P<0.001; χ2=42.750, P<0.001). That the OS [HR=0.44, 95%CI (0.25, 0.77), P=0.004] and the CSS [HR=0.46, 95%CI (0.25, 0.84), P=0.011] in GSCC patients were significantly improved by surgical treatment, showed by the results of multivariate prognostic analysis via Cox proportional hazard mode. Conclusions Surgical treatment was an independent factor affecting the prognosis of GSCC, and it could improve the OS and the CSS. As for the modus operandi, R0 resection should be recommended.