Joint leakage is one of the major problems for shield tunnels. This paper proposes a novel approach to assessing the sealant performance of gasketed joints based on the A-star search algorithm. By treating the contact surface of gaskets as the search space of the optimal leakage path, the water leakage process is modeled as a path optimization task informed by the nonuniform distribution of contact stresses. This new approach determines the optimal leakage path together with the associated leakage threshold. The validity of the proposed algorithm is demonstrated through a comparison with experimental test results, followed by a comprehensive parametric study of joint opening, offset, and self-sealing effect. The results elucidate the dominant role of localized contact in sealant performance, which is notably affected by joint deformation. Scenarios of large offsets and small openings exhibit abnormal increases in leakage thresholds and elongated leakage paths due to additional gasket-to-groove contacts and increased heterogeneity of contact stress distribution. The self-sealing process exerts a pronounced nonlinear impact on the sealant performance of gasketed joints, leading to a rapid escalation of the leakage threshold when the lateral water pressure surpasses 0.5 MPa. Additionally, lateral pressurization induces nonuniform distribution of non-conductive regions, resulting in convoluted optimal leakage paths.