There are many models available for studying sediment yield (SY) and soil erosion (SE) world-wide, although there are not any model which are perfectly physically-based, the mathematical equations representing the individual methods in these models typically take a lot of assumptions and empirical or conceptual methods into account. Input requirements, complexity & capability, temporal & spatial scale, process re-presentation, and output types of these models differ significantly. The main objective of the present study is to scrutinize the physical models for SE and SY with respect to their strengths and mentioned factors. For investigations on SE and SY, the literature advises for the use of models like A.G.N.P.S., W.E.P.P., SWAT, & SHETRAN etc. Mostly existing erosion and yield models can only simulate soil detachment and sediment delivery procedures at the hill-slope scale. The models for channel erosion and reservoir siltation processes have seen less development. The study also offers a set of guidelines for choosing an appropriate model for a particular application/case study. This study will be helpful to understand the principle, structure, application fields and future development trends of physically based model systematically for SE and SY.