Underground utility tunnels are important bases for the development of smart cities. The quality and safety of utility tunnels directly affect the resilience of smart cities. Owing to the complex construction technology involved in building utility tunnels, multiple stakeholders, and asymmetric information, the general contractor can behave opportunistically. To improve supervision efficiency in utility tunnel projects, a supervision model based on prospect theory and evolutionary game theory was developed, and the main factors influencing government and general contractors' strategy choices were investigated. Given the limitation of punishment mechanisms in the prevention of opportunistic behaviours, a government incentive regulation model was further established, and the action mechanism of incentives on opportunistic behaviours under different risk attitudes of the contractors was analysed. The results indicated that the opportunistic behaviours of the general contractor were positively related to safe management, costs, and illegal income and negatively related to accident loss and punishment. The effect of government incentive regulation was influenced by the different risk preferences of general contractors. Punishment and incentive regulations were complementary.