When constructing open tunnel boring machine tunnels in rheological strata, the effects of rock creep can lead to a continuous increase in the support stress, resulting in significant safety risks. This paper conducts a case study based on Xinjiang TBM tunnel project, where field experiments were carried out to obtain data on surrounding rock deformation and support stresses. Utilizing the acquired geological parameters, support parameters, TBM tunneling parameters, and monitored data on surrounding rock and support structures, a data-driven model was developed using machine learning. Furthermore, to provide a visual and accurate assessment of tunnel support safety, the expert authority coefficient (EAC) method was proposed. This method was combined with analytic hierarchy process (AHP) and technique for order of preference by similarity to ideal solution (TOPSIS) to construct a knowledge-driven model. Practical engineering validation has demonstrated that the knowledge-driven model incorporating EAC-AHP with TOPSIS can efficiently and accurately assess the safety of open TBM tunnel. A data and knowledge dual-driven model was proposed by combining the above two models, which can accurately evaluate the safety of the entire tunnel support in the absence of sufficient data and compensate for the shortcomings of traditional methods. The proposed data and knowledge dual-driven model allows for an accurate and efficient evaluation of tunnel support safety. Application of this model in actual case studies indicates its significant practical value.