The invention discloses a support vector regression machine fusion-based expressway road section mean velocity estimation method. The method includes the following steps that: 1) an expressway to be researched is selected, and basic data are obtained; 2) a road section mean velocity is obtained through using a trajectory method-based correction algorithm and according to fee charging data; 3) passenger service vehicles and hazardous article transport vehicles of which the single operating mileage exceeds 800 kilometers are adopted as floating vehicles, a road section mean velocity is obtained through using a speed-time integral model and a weighted average method; and 4) a support vector regression machine fusion algorithm is adopted to fuse the two road section mean velocities, so that a support vector regression machine fusion-based expressway road section mean velocity can be obtained. According to the support vector regression machine fusion-based expressway road section mean velocity estimation method of the invention, expressway road section fee charging data and Beidou satellite positioning system data about the passenger service vehicles and hazardous article transport vehicles of which the single operating mileage exceeds 800 kilometers are combined, the fused road section mean velocity is obtained through the support vector regression machine method, and therefore, the accuracy of velocity estimation is improved, the daily monitoring and emergency disposal of expressway managers can be benefitted, and the operational efficiency of expressways can be effectively improved.