As an application of the Internet of Things, the connected vehicle system (CVS) plays an important rolein smart city. While the vehicles in a CVS inevitably suffer with environmental noise interference, which affect thestability of the whole system. So in this paper, the consensus of autonomous CVS is researched in the presenceof stochastic multiplicative noise. The parameter restriction method is first proposed which achieves the controlrequirements by restricting the parameters of the designed control protocol. Then we optimize this method so thatit can resist the noise with slightly larger intensity. However, this method has limited ability to resist noise oflarge intensity. Next we propose another method that is called the interaction network-based method, which canensure the convergence of the system through introducing a virtual system that interacts with the original CVS. Byselecting parameters of the interaction network properly, such method has a good performance in resisting largerintensity noise. In the mentioned two methods, the convergent results are derived to ensure the stability of theplatoon by using the stochastic differential equation(SDE). Finally, the validity of the proposed methods is verifiedin the designed simulations.