<p>Packet loss has a detrimental effect on the cooperative control performance of connected and automated vehicles, as it results in the loss of critical control information. To address this challenge, this paper presents a stochastic distributed algorithm to ensure cooperative control accuracy and convergence under packet loss conditions. Based on the consensus form derived from the original centralized optimization problem, a distributed computing structure is designed by deriving the analytical solution of the consensus variable and decoupling a specific coefficient matrix. Subsequently, a stochastic distributed algorithm is proposed to numerically solve this computing structure in the context of packet loss. Accordingly, the global convergence of this stochastic distributed algorithm is theoretically analyzed from the perspective of operator analysis with the convex assumption for the feasible domain. The cooperative control, computational efficiency and influence of packet loss on the numerical optimization performance of the proposed algorithm are validated and analyzed through simulation and experimental tests under packet loss conditions in an intersection scenario.</p>