PURPOSE: To review our institutional experience in the surgical treatment of pediatric chronic pancreatitis (CP) and evaluate predictors of long-term pain relief. METHODS: Outcomes of patients =21 years surgically treated for CP in a single institution from 1995 to 2014 were evaluated. RESULTS: Twenty patients underwent surgery for CP at a median of 16.6 years (IQR 10.7-20.6 years). The most common etiology was pancreas divisum (n = 7; 35%). Therapeutic endoscopy was the first-line treatment in 17 cases (85%). Surgical procedures included: longitudinal pancreaticojejunostomy (n = 4, 20%), pancreatectomy (n = 9, 45%), total pancreatectomy with islet autotransplantation (n = 2; 10%), sphincteroplasty (n = 2, 10%) and pseudocyst drainage (n = 3, 15%). At a median follow-up of 5.3 years (IQR 4.2-5.3), twelve patients (63.2%) were pain free and five (26.3%) were insulin dependent. In univariate analysis, previous surgical procedure or >5 endoscopic treatments were associated with a lower likelihood of pain relief (OR 0.06; 95% CI 0.006-0.57; OR 0.07; 95%, CI 0.01-0.89). However, these associations were not present in multivariate analysis. CONCLUSION: In children with CP, the step-up practice including a limited trial of endoscopic interventions followed by surgery tailored to anatomical abnormalities and gene mutation status is effective in ensuring long-term pain relief and preserving pancreatic function.