We　propose　to　use　an　expert-driven　feature　selection　scheme　to　diagnose　and　predict　Large　for　Gestational　Age　(LGA)　fetuses.　A　Fetus　with　excessive　birth　weight　exhibits　adverse　neonatal　and　maternal　complications.　Early　intervention　can　slow　progression　and　prevent　the　upcoming　complication　of　the　disease.　In　this　research,　four　well-known　machine　learning　classifiers　with　ten-folds　cross-validations　are　used　to　authenticate　the　proposed　scheme.　A　Master　feature　vector　is　created,　and　an　expert-driven　feature　selection　scheme　is　proposed,　which　is　later　compared　with　existing　published　researches,　master　feature　file　created,　and　with　an　automated　feature　selection　scheme.　The　best　performance　metrics　(precision　and　AUC)　scores　are　produced　by　random　forest　and　logistic　regression　classifiers　with　the　proposed　expert-driven　feature　selection　scheme.　The　proposed　scheme　played　an　essential　role　in　elevating　prediction　precision　and　AUC　scores　from　0.71　and　0.70　to　(0.9461　and　0.8172)　and　(0.9174　and　0.8281),　respectively.　Therefore,　we　recommend　obstetrician＇s　to　update　the　prognosis　process　for　LGA　identification　using　expert-driven　feature　selection　scheme.　©　2019　IEEE.