This　paper　mainly　proposes　an　optimization　model　for　supermarket　vegetable　sales.　By　establishing　the　distribution　law　of　sales　volume,　the　relationship　between　category　and　item　sales　volume,　and　the　fitting　model　of　the　relationship　between　sales　volume　and　cost　plus　pricing,　the　paper　aims　to　make　replenishment　and　pricing　decisions　for　supermarket,　so　as　to　achieve　better　operating　earnings.　The　graph　of　past　sales　volume　is　drawn　and　the　cyclical　and　seasonal　trends　of　sales　volume　are　found.　Taking　the　month　as　the　classification　unit,　the　distribution　relationship　of　the　sales　volume　of　each　category　and　single　product　with　the　month　is　established.　The　trend　of　merchandise　sales　and　the　impact　of　holidays　on　sales　are　analyzed.　Spearman　coefficient　analysis　was　used　to　study　the　correlation　between　category　and　single　product,　and　the　combination　of　category　and　single　product　with　strong　correlation　was　found.　The　weighted　average　method　was　used　to　evaluate　the　different　contribution　degree　of　each　item　to　its　category,　and　the　daily　selling　price　of　each　item　was　calculated.　Based　on　the　reasonable　assumption,　the　profit　model　of　supermarket　is　established,　and　the　sales　volume　and　markup　rate　are　positively　correlated　with　the　profit　of　supermarket.　The　penalty　function-LSTM　model　is　used　to　fit　the　relationship　between　sales　volume　and　selling　price　to　realize　the　pricing　strategy　of　price　maximization.　This　paper　forecasts　the　daily　replenishment　volume　and　sales　price　from　July　1　to　7.　In　general,　this　paper　comprehensively　uses　statistical　analysis,　weighted　average,　penalty　function-LSTM　model　and　other　methods　to　solve　the　problem　of　supermarket　vegetable　sales,　and　provides　relevant　decision-making　suggestions　and　strategies.　©　2023　IEEE.