This work presents a novel geometric approach to control theory on curved surfaces by establishing a precise connection between isomonodromic deformations of logarithmic connections on principal bundles and control systems on Riemann surfaces. The key insight is that optimal trajectories in control systems can be understood as geodesics with respect to a metric determined by monodromy data. The framework begins by characterizing control systems in terms of logarithmic connections on principal bundles over Riemann surfaces, where the residues at singular points encode the control structure. We establish that controllability is equivalent to the residues generating the full Lie algebra of the structure group under commutator brackets, providing a geometric criterion that can be verified algebraically. The main result demonstrates that optimal control trajectories minimizing a quadratic cost functional correspond precisely to isomonodromic deformations, where the monodromy representation remains constant under parameter variation. This correspondence extends naturally to systems with non-holonomic constraints, where the eigenvalue structure of the residue matrices determines the growth vector of the constraint distribution. The symmetry algebra of the constrained system is shown to be isomorphic to the Lie algebra generated by the residues. We illustrate the theory through a detailed computational example involving principal bundles with special unitary structure group over a hyperbolic surface of genus two.