The most competitive environment generated the need of process performance optimization. Performance optimization means produce the same amount of product, more effectively and spending less money. On the alumina market, that’s fundamental in this economy scenario. Robust Multivariable Predictive Control Technology becomes one of the main tools to optimize this class of plants. This paper will discuss the application and benefits of this technology to alumina digestion units, implemented in 5 interconnected digesters. This digestion interconnection is a whole digestion train, and the plant has 5 of those. The APC philosophy is based on process variability reduction, and consequently operations optimization, against plant constraints. Since alumina – caustic ratio (A/C) is the key plant variable, it has a fundamental role in this variability reduction. The main challenge in this project was to coordinate the use of 5 bauxite grinders and 2 more grinder bauxite flows to the 5 digesters. The implementation was made in 3 phases and the project length was approximately 18 months, generating more than 1.00% increase in overall production, rather than A/C variability reduction.