Admission Control is an important Quality of Service (QoS) mechanism for maintaining performance of streaming media service. The focus of this paper is to analyze the proposed adaptive threshold-based admission control for distributed Video-on-Demand (VoD) systems. A novel distributed VoD system architecture is introduced. To maximize the VoD system's profits, we explicitly classify client requests into K classes and extend the concept of admission control to include the client priority. The expression of the system cost function GAIN is derived by using the M/M/n/n queuing model and the optimal threshold values for different classes are obtained based on the derived formula. Numerical results show that the proposed policy is feasible and performs better than the free threshold policy in terms of the system pay-off.