Localized corrosion is widely recognized as the most common destructive failure mechanism of engineering components in a wide range of industries. The process operates at the microstructural level and is governed by the electrochemical environment and microstructural features of the corroding material, such as grain orientations and grain boundaries. In the presence of mechanical loading, localized corrosion opens hot spots for stress concentration, leading to crack initiation and provoking premature failure, resulting in catastrophic events. Tracking the evolution of the metal-environment interface subjected to the electro-chemo-mechanics effects has been seen as the biggest obstacle in incorporating the role of electrochemistry and microstructural features into mechanistic corrosion models. Recent development in phase-field modeling of corrosion damage [1] has shown that the phase-field method can naturally track the evolution of the metal-environment interface and capture pit-to-crack transition and crack propagation in arbitrary domains without requiring any special treatments or ad hoc criteria. Built upon that formulation, an electro-chemo-mechanical phase-field-based framework for predicting localized corrosion in polycrystalline materials is presented in this work. The model accounts for the synergetic action between the environment and microstructural features and incorporates the mechanical effects in enhancing the corrosion rate. The role of the environment is implemented through the transport of ionic species in the system and corrosion dependence on crystal orientation is incorporated via crystallographic orientation-dependent corrosion potentials. Several 2D and 3D boundary value problems of particular interest are addressed to demonstrate the predictive capabilities of the model. Various microstructures are generated to examine the effect of grain orientation and grain size on localized and mechanically-assisted corrosion. The main findings and discussions will be summarized.