This study proposed an approach for the preliminary risk assessment of slopes with a potential for large volume failure. In this framework, risk is the combined consideration of landslide hazard, exposure, and vulnerability. The level of hazard is classified according to the combination of landslide susceptibility (in space) and activity (in time). The incorporation of landslide susceptibility and activity enables the examination of long-term stability at the site for hazard classification. Several empirical relations were established using regression that incorporated various exceedance probabilities. These relations included (1) that between the landslide source area and the area of slopes prone to large failure (PLF slope), (2) the landslide volume and landslide area, and (3) the equivalent coefficient of friction and the landslide volume. These relations were used together with exceedance probability to determine the area and volume of the landslide source as well as the runout distance. The landslide source was estimated by referring to the slope geomorphology. The influence zone was delineated in reference to the runout distance and the width of the landslide source, and it was partitioned into subzones to distinguish vulnerabilities. With the delineated influenced zone, all the elements at risk (EaR) and their individual vulnerabilities could be identified. This approach classifies the vulnerability level of each EaR according to its location within the affected zone. Each EaR in the affected zone has a different vulnerability weight according to its location. The weighted average of all EaR determines the overall vulnerability of the site. As the end product, potential risks are classified into five levels based on the assessments of hazard, exposure, and vulnerability, which can be easily calculated from the grades of hazard, exposure, and vulnerability. Furthermore, this study demonstrated the applicability of the proposed approach for three PLF slopes in southern Taiwan. The proposed approach to risk classification may serve as a feasible tool for analyzing and ranking the risk level of PLF slopes prior to risk management–related decision-making under conditions of limited available resources.