<p>The complex geological environment poses great challenges to tunnel construction, and effectively avoiding construction accidents is an urgent problem that needs to be solved in current research. To accurately assess the construction risk of tunnels in complex geological environments, based on the uncertainty and fuzziness of accident occurrences, a construction risk early warning model based on improved game theory and a multi-dimensional cloud model is proposed. Firstly, this paper constructs a construction risk evaluation index system with three primary indicators and ten secondary indicators based on the practicality and comprehensiveness of the evaluation indicators. Then, the CRITIC, EWM, and AHP methods are used to calculate subjective and objective weights, and the combined weight is calculated by improving the game theory. Next, cloud theory is used to generate a multi-dimensional cloud model corresponding to each risk level of the predictive indicators, and the construction risk level is determined based on the principle of maximum certainty. Finally, the effectiveness of this model is verified through comparison with other models in engineering cases. The research shows that the risk assessment results based on the improved game theory and multi-dimensional cloud model are effective and reliable. Compared with other models, this model can intuitively and quickly determine the mutual influence between indicators, providing a new method for tunnel construction risk assessment in complex geological environments.</p>