ABSTRACT The development of risk assessment methods for depression and cave-in caused by subsurface cavity in the asphalt pavement is presented. Pavement response models and risk factor concept were used in developing this procedure. A 3D finite element analysis that simulates a subsurface cavity was conducted to generate the synthetic database. Using the generated database, the pavement response models were established to predict critical pavement responses in asphalts pavement with the subsurface cavity using cavity depth/length, asphalt layer thickness, and asphalt modulus. To assess the degree of risk for asphalt pavement with the subsurface cavity, a procedure in determining the risk factor using critical pavement response was proposed in this study. A risk rank system for road cave-in obtained from the Japanese Road Authority was used to verify the risk assessment procedure. It is found that the predicted risk factor using the proposed procedure shows a good agreement with the Japanese risk criteria over a wide range of cavity depth and length. However, the proposed procedure consistently underestimates the degree of risk in asphalt pavements with 1.5–2.5 ratio of normalised cavity depth and greater than 1.5 m of cavity length. This observation demonstrates that the application of the cavity expansion model in this procedure is needed to improve the prediction quality of risk factor.