Selection of highly productive hosts for protein expression is a significant component of bioprocess design. As an alternative to traditional plate, halo, and suppression-based screens, we describe a high-throughput, flow cytometric assay, the Cell Surface Secretion Assay (CeSSA), that can be used to select for improved heterologous protein secretion from a population of S. cerevisiae mutants. A ligand is covalently attached to the cell surface via a PEG linker, and as cells secrete a protein that binds the tethered ligand, the protein is captured on the surface where it can be labeled and the cells sorted using flow cytometry. This report describes three different protein/ligand interactions that have been demonstrated with this system. Single-pass sorting enrichments from 23- to 54-fold have been validated in the separation of a 3-fold higher secretor from a background population of wild-type secretors making this system applicable to large library screening (10(8) clones). A mathematical model was developed to improve the parameters of the assay further. The model was validated with time course data and predicts an optimal screening window. The model also predicts a 60-fold enrichment rate for the validation experiment described above. With the development of this selection system, limitations presented by traditional, particularly plate-based, secretion assays can be overcome so that a larger search space can be examined under conditions closer to the growth physiology experienced by cells in fermentors.