The optimal guidance law (OGL) is based on the assumptions of linear kinematical model of missile guidance and unconstrained control in a linear quadratic regulator formulation. These are strong assumptions that are rarely satisfied in practice; the kinematics of the missile-target engagement is highly nonlinear, and infinite lateral acceleration - which is the control - is a physical impossibility. As a consequence, the application of OGL to the actual nonlinear model shows a large miss distance except for a limited range where the linearity assumption is valid and the control needed is not so large as to hit the constraint boundary. This paper proposes a novel improvisation of the OGL using differential evolution (DE) that compensates the effects of these assumptions and increases the range of validity of OGL. Treating the missile-target kinematics as a black box in which nothing is known except the input and the output, the output is optimized for variations in input using DE. The nominal values of the input variables used for the OGL are used to seed the initial population of trial solutions. ©2007 IEEE.