Abstract: Conventional approaches for analyzing slope dynamic responses to blasting often rely on two−dimensional planar models and indirect vibration loading, which fail to realistically capture the three−dimensional propagation of stress waves and the continuous energy exertion generated by the explosion. To address these fundamental limitations, this study proposed a three−dimensional full−process direct simulation method closely aligned with engineering practice, based on the blasting operation of a quartz mine. The method integrated three−dimensional refined geological modeling, the Jones−Wilkins−Lee (JWL) equation of state for explosives, and the Johnson−Holmquist Ⅱ (JH−2) constitutive model for rock mass damage. This integrated approach enabled a complete simulation spanning from the initial release of explosive energy to the subsequent dynamic responses of the slope, and intuitively reproduced the intricate spatial interactions among blasting stress waves, bench structures, and free surfaces. The simulation results reveal significant elevation amplification effects on the slope under blasting. Dynamic platform responses intensify progressively with increasing elevation. Taking the 2# monitoring line, which is located closest to the blast source and exhibits the highest fitting accuracy, as a representative example, the peak displacement increases dramatically from 0.83 mm to 18.79 mm as elevation increases. Systematic comparisons with field monitoring data further demonstrate that the computational errors of this proposed method are substantially lower than those of traditional two−dimensional profile models. Specifically, the three−dimensional model achieves a Mean Absolute Error (MAE) ranging from 0.20 to 5.66 mm across different monitoring points, which is significantly smaller than the MAE range of 0.92 to 7.56 mm observed for the two−dimensional model. The superior accuracy of the three−dimensional approach confirms that full consideration of spatial stress wave propagation and continuous energy input is essential for reliable dynamic response predictions. This study effectively overcomes the long−standing issues of load simplification and dimensional deficiency inherent in conventional numerical simulations, and provides a high−precision paradigm for dynamic response analysis and stability evaluation of slopes under complex bench terrain conditions.