Abstract: To address the engineering challenges of large slag flow fluctuations, severe furnace bottom air leakage, and unstable discharge temperatures in coal-fired boiler dry slag discharge systems, this study focuses on a 630 MW coal-fired unit and constructs a coupled heat transfer model between slag and cooling air. A feedforward-cascade closed-loop control strategy is proposed. The strategy integrates a binocular vision system to real-time perceive the slag mass flow on the conveyor belt, and designs a slag flow feedforward module based on a retro-compensation mechanism to proactively respond to abrupt slag flow changes. Simultaneously, a conveyor speed feedforward compensation module employing quadratic smoothing attenuation is developed to suppress temperature fluctuations induced by speed variations. To further enhance control performance, key parameters of the feedforward controller are optimized using the Particle Swarm Optimization (PSO) algorithm, with the objective of minimizing the Integral of Time-weighted Absolute Error (ITAE). The proposed strategy is validated on the Matlab/Simulink platform and through 9 hours of field operational data under multiple typical disturbances. Results show that, under setpoint step changes, the rise time of the proposed strategy is reduced by 63.5% compared to conventional PID control. Under disturbances of abrupt slag flow changes, boiler load variations, and furnace negative pressure fluctuations, the maximum slag discharge temperature deviations are limited to 1.3℃, 1.3℃, and 0.2℃, respectively—significantly lower than the 5.1℃, 3.7℃, and 4.8℃ achieved by PID control. Field data demonstrate that the slag discharge temperature setpoint is safely elevated from 140℃ to 148℃ , resulting in a 14.2% reduction in average cooling air volume and a 4℃ increase in furnace inlet air temperature. This study provides an effective technical pathway to resolve operational challenges in dry slag discharge systems, offering significant support for improving boiler thermal efficiency and power plant economic performance.