The invention discloses a method and device for predicting frequently-used routes of trucks and belongs to the field of intelligent traffic big data. The method includes the steps of collecting trajectory data of each truck, performing road matching on the trajectory data in a preset period of time to obtain each corresponding driving route, and trimming each driving route to obtain each corresponding trunk route; dividing each trunk route according to preset conditions to obtain a plurality of sub-route sets, calculating the weight of each sub-route in the plurality of sub-route sets, and determining the frequently-used routes of the trucks in the plurality of sub-route sets respectively. In the method and device for predicting frequently-used routes of trucks, the driving route of each truck is matched based on historical trajectory data of each truck, so that the frequently-used routes of the trucks are mined and predicted, and the predicted frequently-used routes have high experience value and accuracy, thereby providing an experience route for truck drivers and improve the safety of truck travel.