A bus priority signal control method based on variable phase was proposed to adapt to the operation characteristics of intersections with large traffic flow fluctuation. The method allows the phase sequence and phase combination of intersection signal to be dynamically adjusted in real time. Based on the bidirectional and real-time information transmission in Internet vehicles environment, a prediction model of bus arrival time in different situations was developed. The model considers the time and speed of the bus and the queuing status of the intersection when the bus enters the communication range. Then, the green time were determined for different candidate schemes according to bus arrival time. Aiming at the highest passenger pass rate, the optimal scheme was selected from candidate schemes as the next green phase. Finally, the COM interface of VISSIM and Python were used to achieve the proposed method in Internet vehicles environment. The intersection of Fei Yue Road and Guang Gu Street in Changchun was taken as an example. The test shows this method can effectively reduce the per person delay of intersections when the road saturation is 0.8, 0.9 and 1.0. When the road saturation is 0.9 and the bus proportion is 2%, 4% and 6%, the method can reduce the per person delay of buses by more than 40%. Therefore, this method is more suitable for unsaturated intersection with lower proportion of buses.