This paper utilizes the hierarchical analysis method to create an intellectual property protection index model to measure the level and hierarchy of intellectual property protection from the judicial, administrative, and social perspectives and proposes solutions accordingly. From the standpoint of time series prediction, the deep learning theory time series is applied to fully explore the time series characteristics of the flow and a comprehensive assessment of the development status of online intellectual property protection from 2019-2024 is made. According to the findings, pornographic content is the most serious IPR infringement problem in cross-border big data flows, accounting for approximately 38% of all violations. Cases concerning domain name infringement show a gradual upward trend from 2014-2023. The automatic fulfillment index in the judicial protection index is the lowest, at 2.3, and it is important to strengthen this aspect in the future. The overall index of 75.6 in 2019 to 90.1 in 2023 indicates that the trend of intellectual property protection in China is generally positive.