Rooftop photovoltaic (RPV) systems can increase local temperatures and reduce module efficiency due to high solar absorption, whereas photovoltaic-integrated green roofs (PVIGR) can mitigate these thermal effects through their vegetated layers. Given the intricate coupled multiphysics processes governing the PVIGR microclimate, a CFD-based numerical framework is essential for quantifying the influence of meteorological conditions and system configurations, as well as for elucidating the underlying mechanisms under real-world operating conditions. However, existing numerical models often oversimplify physical processes within vegetation and inadequately represent rooftop conditions, which limits mechanistic understanding and, consequently, the development of targeted optimization strategies. This study develops a well-validated transient 3D CFD-based multiphysics modeling framework for the PVIGR microclimate, resolving turbulence, radiative heat exchange, shading, transpiration, and PV energy conversion. Three key methodological choices are evaluated: transient versus steady-state simulations, uniform versus power-law inlet wind profiles, and surface radiation versus heat source methods. Results show that transient simulations avoid exaggerated local overheating, the power-law wind profile better represents rooftop airflow, and volumetric heat sources accurately capture heat transfer within tilted PV modules, while surface radiation for vegetation surfaces accounts for uneven energy input caused by shading. Next, microclimate comparisons show that under the simulated environmental inputs, PVIGR lowers roof and module temperatures by up to 13 °C and 1.6 °C relative to RPV. These temperature reductions correspond to an estimated decrease of up to 80 W/m² in roof surface heat gain and a power increase of up to 2.6 W for a single module under the analyzed conditions. Finally, mechanism analysis shows that while vegetation mainly enhances module radiative heat loss, sub-module convection remains the dominant cooling pathway, suggesting that improved convective transfer may be a promising direction for enhancing PVIGR performance. The proposed numerical framework provides a modeling basis and mechanistic insights that inform PVIGR design and optimization. The results also indicate the potential of PVIGR systems for improving thermal regulation of the roof and PV performance.