Neuroblastoma (NB) is one of the most common solid tumors in children. Glycosyltransferases (GTs) play a crucial role in tumor development and immune escape and have been used as prognostic biomarkers in various tumors. However, the biological functions and prognostic significance of GTs in NB remain poorly understood. The expression data from Gene Expression Omnibus (GEO) and Therapeutically Applicable Research to Generate Effective Treatments (TARGET) were collected as training and testing data. Based on a progression status, differentially expressed GTs were identified. We constructed a GTscore through support vector machine, least absolute shrinkage and selection operator, and Cox regression in NB, which included four prognostic GTs and was an independent prognostic risk factor for NB. Patients in the high GTscore group had an older age, MYCN amplification, advanced International Neuroblastoma Staging System stage, and high risk. Samples with high GTscores revealed high disialoganglioside (GD2) and neuron-specific enolase expression levels. In addition, a lack of immune cell infiltration was observed in the high GTscore group. This GTscore was also associated with the expression of chemokines (CCL2, CXCL9, and CXCL10) and immune checkpoint genes (cytotoxic T-lymphocyte–associated protein 4, granzyme H, and granzyme K). A low GTscore was also linked to an enhanced response to anti–PD-1 immunotherapy in melanoma patients, and one type of tumor was also derived from neuroectodermal cells such as NB. In conclusion, the constructed GTscore revealed the relationship between GT expression and the NB outcome, GD2 phenotype, and immune infiltration and provided novel clues for the prediction of prognosis and immunotherapy response in NB.