The geological conditions of subsea tunnels are extremely complex, and a water inrush disaster is prone to occur when crossing a fault. Based on a geological analysis using cloud theory, analytic hierarchy process (AHP), and entropy weight theory, we aim to establish a normal cloud model with combination weight for water inrush risk evaluation in subsea tunnels. We select seven main factors, namely, the surrounding rock grade, rock integrity, overburden thickness, seawater thickness, tunnel section, permeability coefficient, and construction technology level as risk assessment indicators for the evaluation model. The risk is divided into five levels, and the numerical characters of the cloud model are calculated based on the standard of each risk level. The subjective weight and the objective weight are determined using AHP and the entropy weight method. The combination weight is calculated using the Delphi method. The comprehensive certainty degree is used to determine the risk level of water inrush. The model is applied to the Qingdao subsea tunnel to evaluate the water inrush risk crossing the F_4-1 fault, and the results are basically consistent with the actual excavation conditions. This article provides a new idea for the risk assessment of water inrush in the subsea tunnel.