To meet the needs of managers and travelers to have a deep insight of the traffic status of the highway network, a traffic status discrimination model is constructed based on highway GIS network data and toll data, which is used for discriminating the traffic status of the entire highway network. The road network is divided into road sections based on GIS data and the static features of road sections are extracted. The dynamic features of road sections within a 5-minute time window are extracted based on toll data. Model data are annotated using semi-automatic annotation methods. A dual random forest model is used to address the issue of data imbalance. The experiment shows that the accuracy of the road network traffic status discrimination model proposed in this paper is superior to existing single section traffic status discrimination models.