The relatively high cost of Supercritical Fluid Extraction (SFE) processes has been a major obstacle in the widespread use of the technology. In this work, it is demonstrated that this cost can be reduced significantly by applying the mathematical programming approach and formulating the problem of designing these processes as a Mixed Integer Nonlinear Programming (MINLP) model. Application of the model to the isolation of beta-carotene from algal cells has shown that SFE can easily outperform traditional organic-solvent-based extraction processes. Finally, a new heuristic is proposed to assist in the design of SFE processes which involve poorly soluble natural products.