Here we explore the opportunity to design and then produce tailored release of therapeutic drugs from microcapsules. By use of "building blocks," formed from well characterized microcapsule populations, an inverse design algorithm has been developed that provides an optimal (in a least squares sense) combination of building blocks to achieve a desired release history. Previously we have reported experiments and a well validated mathematical model for computing drug release histories from PLG microcapsules, and these form the backbone of the present optimization algorithm. To expand our available basis for finding useful optimal solutions, we also report work to validate the mathematical model for two different molecular weights. Thus, our building blocks comprise populations that differ by microsphere mean diameter, polydispersity, and polymer molecular weight, giving three separate parameters that effect drug release rate, and from which we build a foundation for our tailored release. Here we have taken a basis of six different microcapsule release systems, from which we build a tailored release history using constrained optimization to fit a prescribed release profile. Comparison of predicted release with measurements from the tailored microcapsule populations was found to produce excellent results, with correlation coefficients greater than 0.98. By way of demonstration, a triple pulse design is described that illustrates the power of the method.