Depending on the type, quality and quantity of available data, different methods exist to determine the hydrocarbon–hydrocarbon contacts in reservoirs. These methods are either direct or indirect. While the direct method involves consciously drilling a well to penetrate potential fluid contacts, the indirect methods include pressure measurements across at least two different fluid types, and/or, the acquisition and subsequent characterization of downhole fluid samples. Recognizing that most petroleum reservoirs are compositionally graded to varying degrees, compositional grading simulation (CGS) is an inexpensive indirect method of evaluating potential gas–oil contact (GOC) in petroleum reservoirs. However, this typically requires detailed pressure–volume–temperature (PVT) data of the hydrocarbon fluids. But, in many cases, PVT data may be limited. To address the challenges in such cases, this paper proposes a new predictive model and practical workflow, hence extending the applicability of CGS. Employing several PVT datasets and rigorous CGS for known black and volatile oil reservoirs, a new semiempirical function is developed for estimating saturation pressure gradient from gas oil ratio (GOR), which is a readily measurable fluid property. Complementing this function, is the identification of a robust empirical model to predict saturation pressure. The associated workflow provides a systematic integration of the predictive saturation gradient and saturation pressure models to estimate the degree of undersaturation, and hence, potential GOC depth in a reservoir characterized by limited PVT dataset. For validation tests, some cases of saturated and undersaturated reservoirs in the Niger Delta are presented. The results confirm the applicability of the new method in the Niger Delta and elsewhere.