Objectives To determine the use of surgical resection of metastatic disease in a large national sample and its association with overall survival. Methods The National Cancer Database was queried for patients with metastatic bladder cancer (2004–2016). Overall survival was assessed using Kaplan–Meier and multivariable Cox analyses. The associations between covariates and use of metastasectomy were assessed with multivariable logistic regression. Results Of the 16 382 patients with metastatic bladder cancer included, 6.8% underwent metastasectomy. Its use increased over time (4.7% in 2004 to 6.6% in 2016; per year odds ratio 1.02, 95% confidence interval 1.00–1.04, P = 0.019). Median survival was 7.0 months for patients who received metastasectomy and 5.1 months for those who did not (hazard ratio 0.85, 95% confidence interval 0.79–0.91, P < 0.001). In subgroup analyses, metastasectomy predicted longer survival in patients with lung (hazard ratio 0.73, 95% confidence interval 0.61–0.88, P = 0.001) or brain metastases (hazard ratio 0.58, 95% confidence interval 0.35–0.96, P = 0.035) and in patients with variant histology (hazard ratio 0.80, 95% confidence interval 0.69–0.93, P = 0.003). Conclusions In a national sample, the use of metastasectomy for bladder cancer is low. Furthermore, metastasectomy is associated with longer survival overall and in multiple subgroups. However, these results should be validated in future studies.