Abstract: To explore the influence of the coupling relationship among unsafe behaviors of motor vehicle drivers on a multi-factor coupled road traffic system, a combined model of the N-K model and the coupling degree model were constructed based on a total of 739 accident data from the National Vehicle Deep Investigation System (NAIS) database from 2017 to 2019. The causes of road traffic accidents are explored through the N-K model. The risk coupling value function of drivers' unsafe driving behaviors was constructed in combination with the coupling degree model to study the coupling effect between unsafe behaviors. The research finds that the level of road traffic risk shows a significant increasing trend with the increase in the number of coupling factors. Among them, the interaction information value T(a,b) of the driver-vehicle two-factor coupling is 0.0167, which is 9.01 times the driver-road interaction information value and is also the combination with the highest risk in the two-factor coupling. The value of the four-factor coupling interaction information reached 0.0220, which indicates the strongest risk effect. Among the unsafe behaviors of drivers, speeding (with a risk coupling value of 0.0822) and failure to yield as required (with a risk coupling value of 0.0683) are significantly higher than other behaviors. The number of accident cases caused by these two behaviors accounts for 43.4% of the total number of cases, and they are the core risk sources causing traffic accidents.