Background: Intravenous cannulation is a common procedure and a valuable skill in the neonatal intensive care unit. Standardized procedure and personnel training are needed in the unit to prevent hospital-acquired blood stream infections. Hence, we evaluated the effect of training using a low-fidelity simulation on the improvement of the aseptic non-touch technique during intravenous cannulation and knowledge retention after 6 months. Methods: The study was conducted in a tertiary care neonatal unit from June 2017 to July 2018. All the staff nurses and junior resident doctors posted in the neonatal intensive care were included in the study. A protocol and checklist score sheet was developed. The score sheet consisted of 23 items with a total score of 46. Participants were expected to obtain a minimum of 80%. A pre-test was conducted initially, followed by a formal training and then a post-test. The NITA newborn venous access mannequin was used to facilitate the training. A re-training for new nurses was conducted after 6 months. Data were analyzed using paired t-test. Results: A total of 29 doctors and nurses were enrolled in the training. The mean pre-test score was 29.93 compared to 42.66 in the post-test scores (mean difference 12.24(95% confidence interval: 9.39–16.05), p < 0.01. The mean scores were higher when the simulation was conducted after 6 months. There was a significant decline in blood stream infection rates from 5.5 to 1.65 per 1000 patient days (p = 0.05). Conclusion: Simulation-based training of health care personnel is a good modality to improve aseptic non-touch technique during intravenous cannulation in the neonatal intensive care unit. Simulation-based training also helps in knowledge retention and standardization of training procedures.