Today, with the progress of medical sciences, increasing the cure probability and survival time is an important goal of cancer treatment. This study compared long-term disease-free survival (DFS) of non-metastatic breast cancer patients based on different molecular subtypes. This retrospective cohort study consisted of 1287 patients diagnosed with breast cancer and treated at Motamed Cancer Institute from 2000 to 2016 and followed up until 2018. Kaplan-Meier curve was fitted to data based on molecular subtypes. Then the semi-parametric mixture cure model was applied to determine the survival and cure probability of molecular subtypes by adjusting clinical and demographic factors. Among 1287 breast cancer patients, 200 (15.5%) cases died. The mean age of patients was 47.00 ± 10.72 years. Women with the HR+/HER2-subtype had the best 5-year survival rate (84.2%), whereas other subtypes had a lower rate as follows: HR+/HER2+ (77.3%), triple-negative (76.5%), and HR-/HER2+ (62.3%). Kaplan-Meier curve calculated a cure rate of about 60% and patients who survived more than 150 months were intuitively considered cured. After adjustment for clinical and demographic variables, the cure probability of HR-/Her2+ patients was substantially lower than HR+/HER2- patients (OR = 0.22), though there were no significant variations in short-term DFS based on molecular subtypes (HR = 0.91). Our results confirm that the most prevalent breast cancer was HR+/HER2- tumor type which had the best prognosis. It is also concluded that HR-/HER2+ patients had the worst outcomes, with the highest rates of recurrence and metastasis and the lowest overall and disease-free survival rates.