There have been many models to estimate reserves and predict oil production performance using the relationship between water cut, fw, (or water-oil ratio, WOR) and cumulative oil production (Np) in the literature. However, it is difficult to choose the suitable models for specific reservoirs. On the other hand, consistency and accuracy are yet to be improved. In this study, several frequently used models for predicting cumulative oil production using water cut have been compared using production data from low permeability reservoirs. These models include the conventional model, the Ershaghi–Omoregie model, the Purvis model, the Arps model, the Bondar–Blasingame model, and the Warren model. All of the models were applied to production data, respectively, and then compared in one single figure, that is, fw versus Np, for one set of production data from both reservoirs and the core sample. To do so, it facilitated the comparison of different models. Otherwise, it may be difficult to make the comparison for all of the models because the models have different dependent variables. The analysis and discussion to the results have been conducted. The results have demonstrated that no model could fit all of the cases studied. Each model has the advantages and limitations. However, the Warren model is better than the other five models statistically. It fits most of the cases studied satisfactorily.