Near-infrared (NIR) light-induced imaging-guided cancer therapy has been studied extensively in recent years. Herein, we report a novel theranostic nanoplatform by modifying polyoxometalate (POM) nanoclusters onto mesoporous silica-coated upconversion nanoparticles (UCNPs), followed by loading doxorubicin (DOX) in the mesopores and coating a folate–chitosan shell onto the surface. In this nanoplatform, the core–shell structured UCNPs (NaYF4:Yb,Er@NaYF4:Yb,Nd) showed special upconverting luminescence (UCL) when irradiated with high-penetration 808 nm NIR light, and the doped Yb and Nd ions endowed the sample with CT imaging properties, thus achieving a dual-mode imaging function. Moreover, the simultaneously generated heat mediated by the 808 nm NIR light may coordinate with the chemotherapy generated from the released DOX to realize an efficient synergistic therapy, verified by diverse in vitro and in vivo assays. The coated folate–chitosan shell can target the platform to tumor tissues when it was transported in the blood vessels and accumulated in tumor sites via the enhanced permeability and retention effect (EPR). Due to the acidic and reductive microenvironment of the tumor, the DOX released quickly with the dissolved folate–chitosan shell, exhibiting obvious tumor microenvironment (TME) responsive properties. The smart imaging-guided therapeutic nanoplatform should be highly promising in TME responsive therapy.