Muscle-invasive bladder cancer (MIBC) is associated with poor predictability of response to cisplatin-based neoadjuvant chemotherapy (NAC). Consequently, the benefit of NAC remains unclear for many patients due to the lack of reliable biomarkers predicting treatment response. In order to identify biomarkers and build an integrated and highly accurate model to predict NAC response, we performed a comprehensive transcriptomic and genomic profiling on tumors from 100 MIBC patients. Our results showed that the expression of the top genes associated with response, as well as the expression of growth factor genes and cell cycle regulators are highly correlated with NAC response. Most importantly, we found a novel signature related to the WNT signaling pathway that alone was highly correlated with NAC response and showed high accuracy in predicting NAC response (AUC=0.76). Additionally, mutations in the DNAH family genes (DNAH8, DNAH6 and DNAH10) and deletion in KDM6A were also highly correlated with NAC response. Using our comprehensive molecular analysis as a backbone, we developed two machine learning (ML) models, one incorporating both transcriptomic and genomic features (RF-RW), and the other using only transcriptomic data (RF-R). Both models demonstrated promising performance (AUC=0.82) as predictive models of response to NAC in MIBC. RF-RW and RF-R, after external validation, could potentially change the management of MIBC patients by selecting ideal candidates for NAC.