Device-induced thrombosis remains a major clinical challenge for mechanical heart valves, driven by complex interactions between unsteady hemodynamics, leaflet motion, and blood–surface interactions. While previous studies have extensively investigated thrombogenic risk using hydrodynamic surrogates such as wall shear stress (WSS), oscillatory shear index (OSI), and platelet activation models [1,2], direct numerical prediction of thrombus deposition and growth remains limited [3,4]. Existing approaches are largely based on Eulerian CFD frameworks, or phenomenological models, with comparatively few studies addressing thrombus formation within a fully Lagrangian smoothed particle hydrodynamics (SPH) context. This study addresses this gap by proposing a novel multiscale, multiphysics SPH-based framework that couples fluid–structure interaction (FSI) of a bi-leaflet mechanical heart valve with an explicit thrombus formation and agglomeration model. Building upon validated SPH formulations for cardiovascular flows [5,6], the model integrates shear-driven platelet activation, deposition, and growth with valve kinematics in a patient-specific aortic geometry reconstructed from 4D MRI. Six clinically relevant valve mounting orientations are systematically analysed to assess their influence on flow structures, WSS, thrombus deposition patterns, and leaflet dynamics. The results demonstrate that thrombus formation is highly sensitive to local flow separation, low-shear recirculation zones, and asymmetric jet impingement, particularly near hinge and valve mounting regions. Among the configurations studied, a 90° rotated valve orientation exhibits the lowest thrombus burden, most uniform shear distribution, and preserved leaflet mobility, while tilted and vertically displaced orientations show significantly higher thrombogenic risk and, in some cases, leaflet immobilisation. A particle-based safety matrix is proposed to enable rapid comparative assessment of thrombogenic performance across configurations. This work advances the state of the art by demonstrating, for the first time, an SPH-based predictive framework capable of modelling both FSI and thrombus growth in mechanical heart valves. The findings highlight the importance of valve orientation as a purely geometric mitigation strategy and establish SPH as a powerful tool for thrombosis-focused, patient-specific in-silico valve assessment.