Congestion in networks greatly reduces efficiency and pr oduction. Systems with limited capacities or resources require analysis in order to ensure optimal results, which may be in terms of cost, data transmitted or vehicles discharged. A network is like an interconnected web and if one section or link is not performing optimally, the network may not be operating efficiently. The objective of th is research is to maximize link traffic throughput in the long run to alleviate congestion. The approach is to mode l the changes in link traffic st ates as a discrete Markov chain, from mathematical theory, due to its random or stochastic nature. This link traffic flow model can be used for varying stochastic processes with corresponding performance measures, for example, a service rate. While this paper focuses on vehicular traffic flow, other disciplines are invited to collaborate on the study of link traffic flow in similar stochastic systems. This research presents a n ovel blocking probability distri bution to account for congestion based on the M/G/c/c state-dependent queuing model. The objective function with the blocking probability was optimized and the results were compared with a simulation model. The optimal solution to the objective function is a flow at which throughput on the link is maximized for the long run. Under a Vehicle infrastructure Integration (VII) scenario, this model may serve as the basis of link flow control in an effort to achieve the maximum link through in the long run.