Purpose: A nomogram is a reliable tool to generate individualized risk prediction by combining prognostic factors. We aimed to construct a nomogram for predicting the survival in patients with non-metastatic human epidermal growth factor receptor 2 (HER2) positive breast cancer in a prospective cohort. Methods: We analyzed 1304 consecutive patients who were diagnosed with non-metastatic HER2 positive breast cancer between January 2008 and December 2016 in our institution. Independent prognostic factors were identified to build a nomogram using the COX proportional hazard regression model. The prediction of the nomogram was evaluated by concordance index (C-index), calibration and subgroup analysis. External validation was performed in a cohort of 6379 patients from the Surveillance, Epidemiology, and End Results (SEER) database. Results: Through the COX proportional hazard regression model, five independent prognostic factors were identified. The nomogram predicting overall survival achieved a C-index of 0.78 in the training cohort and 0.74 in the SEER cohort. The calibration plot displayed favorable accordance between the nomogram prediction and the actual observation for 3-year overall survival in both cohorts. The quartiles of the nomogram score classified patients into subgroups with distinct overall survival. Conclusion: We developed and validated a novel nomogram for predicting overall survival in patients with non-metastatic HER2 positive breast cancer, which presented a favorable discrimination ability. This model may assist clinical decision making and patient–clinician communication in clinical practice.