Objective: To investigate the impact of pre-treatment TPMT and NUDT15 genotyping on medication selection, tolerability and discontinuation rates of azathioprine or 6-mercaptopurine therapy in children with inflammatory bowel disease (IBD). Methods: A retrospective cohort study was conducted on 181 children with IBD who were scheduled for azathioprine or 6-mercaptopurine therapy at the Department of Gastroenterology, Children's Hospital, Zhejiang University School of Medicine between January 2010 and January 2023. Among them, 168 children who received treatment were divided into a genotyped group and non-genotyped group based on pre-treatment TPMT and NUDT15 genotyping. The incidence of drug-related adverse reactions was compared between the two groups. The impact of genotyping on medication selection and discontinuation rates was analyzed. Chi-square test or Fisher exact test were used for intergroup comparisons. Logistic regression analysis was used to control the confounding factors. Firth Logistic regression analysis was applied for data with complete separation. The probability of discontinuation was assessed using survival analysis with Cox proportional hazards modeling. Results: Among the 181 children with IBD, 13 did not receive azathioprine or 6-mercaptopurine due to genetic variants, while the remaining 168 underwent the therapy (154 cases of Crohn's disease and 14 cases ulcerative colitis; 108 males and 60 females). Excluding the 13 untreated cases, 77 children underwent TPMT and NUDT15 genotyping were assigned to the genotyped group, and the remaining 91 to the non-genotyped group. Adverse reactions included myelosupression (26 cases,15.5%), hepatotoxicity (18 cases,10.7%), gastrointestinal disturbance (25 cases,14.9%), alopecia (12 cases,7.1%), fever (3 cases,1.8%), rash (2 cases,1.2%), and pancreatitis (1 case,0.6%). The incidence of overall adverse reactions was significantly higher in the non-genotyped group compared to that of the genotyped group (40.7% (37/91) vs. 26.0% (20/77), P<0.05). Specifically, the non-genotyped group had a higher rate of gastrointestinal reactions compared to the genotyped group (24.2% (22/91) vs. 3.3% (3/77), P<0.01). Cox regression analysis revealed that non-genotyped group had a higher risk of treatment discontinuation due to the adverse reactions (HR=1.47, 95%CI 0.65-3.30). Conclusion: Pre-treatment genotyping of TPMT and NUDT15 variants can help guide the selection of clinical drugs, reduce the incidence of drug-related adverse reactions and enhance tolerability of azathioprine or 6-mercaptopurine therapy in IBD children.