<p>Riverbank failures are increasing prevalent in tropical flood-prone regions due to intense rainfall, steep terrain, fluctuating river levels, and seismic activity. Recent flash flood events have demonstrated how rapid drawdown can trigger slope instability, particularly in homogeneous sandy riverbanks. These failures are exacerbated by excess pore pressure and lack of reinforcement, particularly in seismic areas. However, numerical studies that address the combined hydraulic and seismic effects, along with reinforcement measures, remain limited. This study investigates riverbank stability along a river in Bengkulu City, Indonesia, under a continuous sequence of hydrological loadings and pseudo-static seismic loading using finite element modelling. The subsoil stratigraphy was spatially modelled using Deep Neural Network (DNN) to predict shear wave velocity (<i>V</i>_<i>s</i>) values from microtremor measurements, accurately defining the critical silty sand overburden. A fully coupled flow-deformation analysis was utilised to capture transient pore water pressures and evaluate the factor of safety (FS), displacement, and failure mechanisms of natural and gabion-reinforced slopes. Results indicate that rapid drawdown reduces the FS of natural slopes by up to 49%, triggering severe instability. Furthermore, applying seismic forces during the post-drawdown recovery phase, where residual excess pore water pressures persisted, caused additional FS reductions of up to 48%. Conversely, gabion reinforcement increased FS by 29%, successfully prevented failure, decreased displacement, and altered failure surfaces into deeper, more stable rotational zones. These findings provide critical insights into time-dependent riverbank failure mechanisms and demonstrate the effectiveness of gabion reinforcement under coupled hydrological and seismic loads.</p>