Background Bladder cancer (BC) is a heterogeneous disease characterized by high recurrence and a poor prognosis. Molecular subtypes of BC portend personalized and precision medicine. However, whether there is a difference in therapeutic response to chemotherapy or immunotherapy between different molecular subtypes of BC has not been systematically evaluated.Methods A comprehensive literature search was performed up to October 2020. Consensus clusters 1 (CC1), CC2 and CC3 molecular subtypes were defined according to the heterogeneity and similarity of BC molecular subtypes from published studies to perform meta-analysis. Pooled odds ratios (ORs) with 95% confidence intervals (CIs) were used to assess the therapeutic response, and a fixed- or random-effects model was used according to the existence of heterogeneity.Results Eight studies involving 1463 patients were included in this research. For immunotherapy, CC3 showed the highest respond rate (CC3 vs CC1: OR 0.52, CI [0.34–0.78], P = 0.002. CC3 vs CC2: OR 0.42, CI [0.28–0.62], P < 0.0001), which was mainly reflected in the highest response rate to atezolizumab (CC3 vs CC1: OR 0.47, CI [0.29–0.75], P = 0.002. CC3 vs CC2: OR 0.38, CI [0.24–0.59], P < 0.0001), and the response rates to nivolumab showed no advantage over CC1 and CC2. No difference in response to the two immunotherapies between CC1 and CC2. For chemotherapy, CC3 had the lowest response rate to the overall chemotherapy (CC3 vs CC1: OR 2.28, CI [1.39–3.74], P = 0.001. CC3 vs CC2: OR = 2.25, 95% CI 1.34–3.76, P = 0.002). Compared with CC2, CC3 responded poorly to both neo-adjuvant chemotherapy (NAC) (OR 1.93, CI [1.09–3.41], P = 0.02) and chemoradiation therapy (CRT) (OR 4.53, CI [1.26–16.27], P = 0.02). Compared with CC1, CC3 only showed a poorer response to CRT (OR 6.07, CI [1.87–19.71], P = 0.003), and no difference in NAC. No difference between CC1 and CC2 subtypes in the response rates to NAC and CRT.Conclusions Our study suggested that molecular classifications are important predictors of cancer treatment outcomes of BC patients and could identify subgroup patients who are most likely to benefit from specific cancer treatments.