Currently, the prognosis of osteosarcoma (OS) remains discouraging, especially in elderly/metastatic OS patients. By impairing the antitumor effect of immune cells, tumor immune microenvironment (TIME) provides an environment conducive to tumor proliferation, which highly requires accelerated nicotinamide adenine dinucleotide (NAD+) metabolism for energy. Recently, many genes involved in the sustained production of NAD+ in malignant tumors have been verified to be possible prognostic indicators and therapeutic targets. Therefore, the current study was to probe into the association of NAD+ metabolism-related genes with TIME, immunotherapeutic response, and prognosis in OS. All OS data for the study were acquired from TARGET and GEO databases. In bioinformatics analysis, we performed Cox analysis, consensus clustering, principal component analysis, t-distributed stochastic neighbor embedding, uniform manifold approximation and projection, gene set enrichment analysis, gene set variation analysis, Lasso analysis, survival and ROC curves, nomogram, immune-related analysis, drug sensitivity analysis, and single-cell RNA sequencing (scRNA-seq) analysis. Cell transfection assay, RT-qPCR, western blot analysis, as well as cell wound healing, migration, and invasion assays were performed in vitro. Bioinformatics analysis identified A&B clusters and six NAD+ metabolism-related differentially expressed genes, constructed risk model and nomogram, and performed immune-related analysis, drug susceptibility analysis, and scRNA-seq analysis to inform the clinical treatment framework. In vitro experiment revealed that CBS and INPP1 can promote migration, proliferation as well as invasion of OS cells through TGF-β1/Smad2/3 pathway. Based on bioinformatics analysis and in vitro validation, this study confirmed that NAD+ metabolism affects TIME to suggest the prognosis of OS.