Considering the cooperative adaptive cruise control(CACC) problem of uncertain connected vehicle systems with heterogonous communication, this paper presents a design method of robust controllers for the connected vehicle system. The Bernoulli stochastic process and tracking model with varying input-delays are integrated to formulate the connected vehicle system with parameter uncertainty, losing packet and time delays. To reduce the design complexity of CACC controllers, the decentralized output feedback control structure and linear matrix inequality technology are used to compute the CACC controller of the uncertain connected vehicle systems with heterogonous communication. Moreover, the results on the stability of the closed-loop system and the string stability of the connected vehicle system are established by the methods of time-delay systems and frequency domain analysis. Finally, a connected vehicle system consisting of four cars with communicated by three channels is used to simulate and illustrate the effectiveness of the proposed algorithms.