The fundamental principles employed to rationally direct biological processes have evolved primarily based on results from trial-and-error experiments guided by scientific intuition. However, the inherent complexity of the intracellular signaling events that drive these processes hinders the ability of intuition to efficiently design experiments for obtaining the desired response. There is a critical need to rationalize the design of experiments using quantitative, model-based approaches. The work presented herein aims to address this need by establishing a control-theoretic approach, utilizing both theoretical and experimental components, to facilitate the design of experimental strategies to predictably direct biological processes. Adaptive model predictive control strategies are paired with nonlinear and sparse grid-based optimization approaches to address applications ranging from the control of cellular differentiation to the scheduled dosing of pharmaceuticals. The developed control algorithms were designed to account for the highly uncertain models and practical experimental limitations characteristic to biological systems. However, despite the biological context, these algorithms address challenges which exist for the control any uncertain system and are expected to be broadly applicable.