The　modeling　samples　for　difficulty　to　measure　operation　indexes　and　abnormal　faults　of　complex　industrial　processes　usually　have　the　characteristics　of　sparse　quantity,　unbalanced　distribution,　and　lack　of　connotation　mechanism　knowledge.　Virtual　sample　generation　(VSG)　is　a　technology　to　expand　the　space　and　quantity　of　modeling　samples　and　has　become　one　of　the　main　ways　to　solve　the　formerly　mentioned　difficulties.　However,　there　are　still　some　problems　in　the　existing　research　results,　such　as　the　lack　of　theoretical　support,　the　unclear　of　category　criterion　and　the　application　boundary.　First,　the　existing　problems　for　difficulty　to　measure　operational　indexes　and　abnormal　fault　modeling　of　complex　industrial　processes　are　described.　The　definition　of　virtual　samples　and　the　connotation　of　virtual　samples　are　combed,　and　the　VSG　implementation　process　for　the　regression　and　classification　problems　is　provided.　Second,　the　research　status　is　summarized　from　the　sample　coverage　area,　implementation　process,　and　application.　Third,　further　research　direction　is　analyzed　and　discussed.　Finally,　the　summary　and　future　challenges　are　given　out.　©　2024　Science　Press.　All　rights　reserved.