Frost formation is a common phenomenon observed across many different fields, including refrigeration, construction, and natural gas processing. However, due to its complex nature, developing an accurate and reliable numerical model remains a significant challenge. Despite previous efforts to tackle this issue, current models still have certain limitations. This paper introduces a modified numerical model for frost formation, developed based on the fundamental mechanisms underlying frost formation. The model utilizes an Eulerian multiphase flow approach coupled with the Lee phase change model. In addition, the approach for determining the maximum frost volume fraction is updated, enabling the model to consider density variations during the frosting process. The model is rigorously validated by comparing it with experimental data on thickness, density, and distribution from various studies. The results indicate that the mean absolute relative deviation (MARD) for frost thickness is 8.97%, while the MARD for density is 16.06%. Furthermore, the frost morphology predicted by the model closely matches the experimental observations reported in the reference.