The invention discloses an urban road network dynamic traffic jam prediction method based on floating vehicle data. The urban road network dynamic traffic jam prediction method based on the floating vehicle data includes the following steps that first, a road section jam state discrimination index is determined; second, the floating vehicle data are converted into traffic flow parameters; third, the road section traffic flow state is judged; fourth, the dynamic traffic capacity C1 of a target road section 1 is predicted; fifth, the dynamic traffic demand D1 of the target road section 1 is predicted; sixth, the jam probability, the jam (if happening) degree LOC and the jam formation time of the target road section 1 are predicted. The urban road network dynamic traffic jam prediction method based on the floating vehicle data is a method for predicting dynamic traffic jams of an urban road network, provides a decision making basis for urban traffic management and control measures and has practical promotional value.