Abstract: The arrival of the 5G era makes the vehicular wireless network more intelligent and faster to realize the communication and interconnection between people, vehicles and things, so as to enhance the safety warning of vehicle driving, fast media access, and improve the driving experience. Aiming at the shortcoming of low spectrum allocation efficiency and slow speed in traditional cognitive vehicular networks, we propose a spectrum allocation algorithm based on improved sparrow search algorithm. The refracted opposition-learning was used to construct the initial population for promoting the population diversity. The sine and cosine optimization, the inertia weight and Cauchy chaotic mutation mechanism were used to improve the optimization precision and speed of standard sparrow search algorithm. The spectrum allocation variables were mapped as the location of sparrow individuals. The network throughput and the access fairness were regarded as the fitness function for evaluating the quality of sparrow location. The improved sparrow search algorithm was used to iteratively search the solution of the spectrum allocation. Numerical simulation results show that the improved algorithm can not only get faster spectrum allocation scheme, and higher-yielding car users, can also guarantee allocation fairness.