To optimize the geometric design of dedicated freight corridors on highways and reduce the probability of truck accidents, PC-Crash software was used to conduct accident simulation tests with speed, horizontal curve radius, longitudinal slope, superelevation slope, hard shoulder width, widen value of curve, adhesion coefficient, and vehicle type (heavy trucks and articulated trains) as risk factors. The significant risk factors were screened by binary Logistic regression analysis, and the prediction model of roadside accident probability on dedicated freight corridors was constructed. Based on the probabilistic prediction results and the thought of "cumulative frequency curve method", the probability thresholds of potential blackspot and frequent blackspot of roadside accidents were given, and the identification method of roadside accident blackspot on dedicated freight corridors was proposed. In view of the sections with frequent occurrence of roadside accidents on dedicated freight corridors, the optimization geometric design of dedicated freight corridors was carried out by using classification and regression tree (CART) algorithm. Results show that the importance of significant risk factors on roadside accidents from the largest to the smallest was as followings: speed, horizontal curve radius, adhesion coefficient, vehicle type, hard shoulder width, longitudinal slope, and superelevation slope. For the dedicated freight corridor section with horizontal curve radius of R≤400 m and operating speed of 60 km/h＜v≤80 km/h, the longitudinal slope should be ＜4%; if this design requirements were not met, the superelevation slope should be ≥4%. For the section of dedicated freight corridor with operating speed of 80 km/h＜v≤100 km/h and horizontal curve radius of 400 m＜R≤800 m, the hard shoulder width should be ≥2.25 m. Additionally, increasing the hard shoulder width had a significant effect on reducing the risk of roadside accidents for heavy trucks.