The supercritical carbon dioxide (SCO2) Brayton cycle system is an energy conversion system with great potential, and its efficiency can be improved significantly by using low-leakage dry gas seals in SCO2compressors and gas turbines. However, the high-pressure and high-speed conditions of a SCO2dry gas seal make the impact of turbulence not negligible. In this paper, the influences of flow state on flow field distribution and sealing performance of a SCO2dry gas seal are studied, the criteria for determining the overall flow state into the sealing gap are proposed. Explicit expressions of the flow factor for flow state determination are obtained based on the artificial neural network algorithm. The results show that turbulent flow is easy to occur in a SCO2dry gas seal, leading to a remarkable error in predicting sealing performance using a laminar flow model. When employing a laminar flow model, there is a large error on opening force in the transition regions between turbulent and transitional flows, and a large error on leakage rate at high inlet pressure and close to the critical temperature compared to those obtained by using a turbulent flow model. In addition, the flow state judgment method proposed in this paper can accurately determine the overall flow state of a SCO2dry gas seal under different operating conditions.