<p>The digestion process is one of the most complex stages in industrial alumina production. Precise control of the caustic ratio, the core parameter governing this process, is critical for digestion efficiency and product quality, yet faces two significant challenges. First, the highly complex mechanism involves multiple coupled chemical reactions, complicating real-time internal state perception. Second, practical implementation is constrained by limited edge computing resources hindering complex algorithm deployment and safety-imposed restrictions on the adjustment range and frequency of control valves, posing significant challenges to the feasibility design of advanced control schemes. To address these issues, a macro-micro synergistic prediction model for caustic ratio in the alumina digestion process is established and a highly feasible edge control strategy is proposed, which takes both deployability and safety into account. Firstly, particle dynamic mechanistic analysis is conducted at the micro-scale, introducing a real-time solid-phase aluminum concentration factor to establish an accurate reaction kinetics model of alumina digestion process. Subsequently, a full-process multi-reaction decoupling analysis is performed at the macro-scale, enabling the construction of a macro-micro synergistic full-process prediction model. Furthermore, an easily deployable edge control architecture is established, integrating an actuator-protective model predictive control method to form the industrial edge control scheme. Finally, a semi-physical simulation platform for the alumina digestion process is established for validation. Experimental results demonstrate enhanced modeling and control accuracy while confirming significant industrial feasibility.</p>