Wire-arc additive manufacturing (WAAM) simulations are critical for predicting material properties, layer morphology, and defect formation, thereby enabling production optimization and reducing the need for costly experiments. However, the complex multiphysics and multiscale nature of WAAM makes the process simulation computationally expensive [1]. Consequently, existing studies focus either on simplified thermo-fluidic behavior for layer morphology or simplified thermo-mechanical analysis for residual stress and distortion. However, neither captures the interlink between layer morphology and residual stress [2]. Moreover, emerging fabrication strategies, such as non-planar deposition, require more versatile modeling approaches to accurately represent curved layers with variable thickness. To address these limitations, we propose a novel, part-scale additive manufacturing (AM) simulation based on the hybrid particle-grid Material Point Method (MPM) [3], which efficiently handles large deformations and complex material behavior. In this framework, material deposition is represented by particles carrying state-dependent properties, which are projected onto an Eulerian mesh to solve the transient heat transfer problem, thereby capturing heating and cooling cycles and providing each particle’s thermal history. This thermal history governs the material state and the associated constitutive response by a temperature-dependent shear modulus, enabling viscous flow at high temperatures and elastic behavior at lower temperatures. After each heating–cooling cycle, a quasi-static mechanical equilibrium problem is solved using thermally induced particle-level strains to compute the evolving distortion and residual stress. The Eulerian grid computes temperature, rate of deformation, velocity, and displacement, which are interpolated to and from particles, ensuring smooth phase transitions and stable particle advection. This AM simulation approach is used to simulate different deposition strategies on thin-walled structures, including non-planar deposition. A parametric study examines the effects of process parameters, such as deposition rate, heat input, interlayer time, and tool path, on layer morphology and residual stress development across various deposition strategies. The proposed multiphysics and hybrid material discretization method strikes a practical balance between accuracy and computational efficiency.