Yield　optimization　aims　to　find　high-yield　designs　of　microwave　components　under　the　uncertainties　in　the　production　process.　However　the　Monte　Carlo-based　yield　optimization　requires　extensive　electromagnetic　(EM)　simulations.　In　this　paper,　the　polynomial　chaos　(PC)　expansion　is　used　to　construct　the　surrogate　model　of　the　EM　responses.　By　combining　the　PC　coefficients　with　the　objective　function,　the　objective　function　is　related　to　the　optimization　variable.　This　approach　allows　for　reliable　yields　and　uses　fewer　EM　simulations　compared　to　the　Monte　Carlo-based　yield　optimization　method.　©　2023　Applied　Computational　Electromagnetics　Society　(ACES).