提出一种基于半马尔可夫网络流量模型的延迟估计算法,用于估计传输链路的延迟超过某个阈值的概率,该值可做为控制器设计及控制策略选择的依据。利用半马尔可夫理论对网络流量进行建模,建立了延迟与网络传输中的队列队长、及队长与网络流速率的函数关系,之后经过推导和有限分割,得出网络延迟概率计算公式,并分析了该算法的复杂度。在计算布朗运动积分时采用It积分实现。仿真证明该算法是有效、可行的。 A delay estimation algorithm based on semi-Markov network traffic model is proposed to estimate the probability of transmission link delay exceeding a certain threshold, which can be used as the basis for controller design and control strategy selection. Semi-Markov theory is used to model the network traffic, and the relationship between delay and queue captain and the captain’s network flow rate is established. Then, the derivation and finite segmentation are used to calculate the network delay probability. The complexity of this algorithm is analyzed. In the calculation of Brownian integral integral it. Simulation shows that the algorithm is effective and feasible.