Enterprises and organizations are increasingly relying on data centers to centrally manage and analyze data. Due to their inherent limitations, traditional data processing methods are difficult to meet the needs of real-time analysis and monitoring. Therefore, this article introduces machine learning for optimization and improvement. This article first discusses the architecture and composition of the data center and then uses the recurrent neural network (RNN) to analyze its application in real-time analysis and monitoring of the data center. Subsequently, based on the specific instructions of the hidden status update and output calculation formula, this paper designs an RNN-based real-time analysis and monitoring model of a data center. Through two sets of simulation experiments, the following conclusions are drawn. Compared with the static modeling method, the overall average reduction in the delay index of the real-time analysis and monitoring model of the data center based on machine learning is about 6.9 ms.