The presented work proposes a unified framework for the integrated design of solvent molecules and process systems. The employed design philosophy allows the identification of solvent molecules based on process performance criteria. It employs multi-objective optimization technology in order to capture the manifold trends and trade-offs characterizing the solvent design space, while avoiding the introduction of unnecessary biases or user defined assumptions. The obtained solvent design information is effectively incorporated into the process synthesis stage through the use of data mining techniques in the form of clustering. The process synthesis framework is sufficiently flexible to accommodate for separation or reactive-separation superstructures of the most general type. The presented method is illustrated through examples in the design of solvents for liquid-liquid extraction, gas-absorption, extractive distillation and extractive fermentation processes.