Addressing industry challenges in liquid cooling systems for high heat-flux devices such as computing servers and high-power power electronic components—including poor heat transfer uniformity, high energy consumption due to flow pressure drop, limited adaptability to dynamic thermal loads, and significant discrepancies between traditional single-phase fluid simulations and practical engineering requirements—this study utilizes the ANSYS simulation platform combined with Fluent numerical fluid analysis and CFD-DEM solid-liquid coupled simulation techniques to investigate multi-objective optimization and intelligent dynamic control strategies for microchannel liquid-cooling plate structures. Using rectangular microchannel plates as the research model, core structural parameters—including channel width, spacing, and height—are optimized while evaluating multidimensional performance metrics such as maximum surface temperature, temperature variation range, and coolant flow pressure drop. An improved NSGA-III multi-objective genetic algorithm is employed to achieve global Pareto optimization of structural parameters, resolving the trade-off between cooling efficiency and flow energy consumption inherent in conventional single-parameter optimization methods. Additionally, to address issues like delayed dynamic response and suboptimal energy efficiency in traditional fixed-flow/temperature operation modes, an adaptive intelligent control system based on fuzzy PID control is developed, enabling real-time precision regulation of coolant flow rates under varying thermal load conditions. Simulation results demonstrate that multi-objective optimization significantly enhances the overall performance of the liquid-cooled plate: the maximum equipment temperature is reduced by 7.2 °C, the temperature range is decreased by 4.5 °C, the flow pressure drop is lowered by 18.3%, and the heat transfer coefficient increases by 8.0%. Under dynamic thermal load conditions with varying gradients, the designed intelligent control system maintains temperature fluctuations within ±1.2 °C; compared to traditional constant-speed operation modes, the system achieves a comprehensive cooling energy efficiency improvement of 12.6%. The CFD-DEM solid-liquid coupling simulation method employed in this study effectively addresses the limitation of conventional single-phase CFD simulations in neglecting particle-induced heat transfer interactions, substantially enhancing numerical simulation accuracy. The proposed integrated design approach combining "static multi-objective optimization of structure + dynamic intelligent system control" provides a robust theoretical foundation and technical reference for developing efficient, low-energy-consumption, and intelligent thermal management systems for high heat flux electronic devices.