作为数字媒体网络视频通信的主要方式,VBR MPEG视频流量的预测能力是直接关系缓冲区设计、动态带宽分配及拥塞控制等提高网络服务质量的关键因素.因此针对MPEG视频流的复杂特性,充分利用人工智能方法的优势,提出并建立了基于模糊神经网络的智能集成VBR MPEG视频流量预测模型.采用模糊预测模型提高预测精度,利用神经网络解决预测的实时性问题.实验结果表明,与标准AR预测模型相比,该模型预测的准确度和可靠性显著提高,且算法简单易于推广到其他方法中使用. As the main way of video communication in digital media networks, the prediction ability of VBR MPEG video traffic is a key factor to improve network service quality which is directly related to buffer design, dynamic bandwidth allocation and congestion control, etc. Therefore, taking full advantage of the complex features of MPEG video streaming Artificial intelligence method, this paper proposes and establishes a intelligent integrated VBR MPEG video traffic prediction model based on fuzzy neural network, which uses fuzzy prediction model to improve the prediction accuracy and neural network to solve the real-time prediction problem.The experimental results show that, compared with the standard AR prediction Compared with the model, the accuracy and reliability of the model are significantly improved, and the algorithm is easy to generalize to other methods.