Introduction: We used road crashes between vehicles and two-wheelers from Yinzhou District Ningbo in 2011–2015 from the China In-depth Accident Study (CIDAS) as sample cases. The risk factors of different injury severity grades were analyzed. Method: The classification tree model was used to screen the possible interaction items, and the corresponding regression model was constructed according to the results of the tree model to explore the influencing factors of cyclist injury. Results: The road types, weather types, gender, age of the riders, and vehicle speed were significantly correlated with the dependent variables. The interaction effect of gender*road type, weather*age, weather*speed and speed*age were obtained through a tree model. Conclusions: The risk of male casualties at the crossroads was 3.31 times higher than that of female casualties at the straight roads. On sunny days, the risk of crash casualties in Ningbo was low, and the fatality risk when the speed reached 38 km/h was 10%. Under the interaction effect of weather and age, the fatality risk in cloudy/foggy and rainy days was almost coincident, and the serious risk in cloudy/foggy conditions was the highest. Practical applications: Through factor analysis, it is confirmed that there is interaction effect among the factors, and it can provide reference for relevant departments to formulate more targeted and effective governance strategies.