Abstract: 【Purpose/significance】By conducting research on patent information security identification, it can help ensure the confi⁃ dentiality of patent information, prevent theft and tampering, and thus protect the rights of innovators. In order to fully protect the rights and interests of patent buyers and provide strong support for patent and intellectual property protection, a patent information se⁃ curity identification method based on blockchain and hybrid representation learning is proposed.【Method/process】We use block⁃ chain technology to build a patent information traceability model, and combine the results of information traceability to build a patent information security identification framework based on hybrid representation learning. Under this framework, convolutional neural net⁃ works and graph convolutional neural networks are used to represent patent information texts and structural security features, forming feature vectors. Using feature vectors as inputs to a fully connected neural network to obtain patent information security prediction re⁃ sults. If there are security risks in patent documents, connect the traceability barcode and traceability query blockchain to query patent historical transaction information, confirm whether the patent documents have been tampered with by matching digital signatures, and use traceability records as a reliable basis to achieve patent information security identification.【Result/conclusion】It was found that this method has strong practicality in patent information security identification work and can significantly improve the effectiveness of patent information security identification.【Innovation/limitation】Applying blockchain and hybrid representation learning to patent in⁃ formation security identification can increase identification accuracy while ensuring information security. Hybrid representation learn⁃ ing methods typically require larger computational resources and storage space, especially for complex patent information identifica⁃ tion tasks, which can lead to increased costs during implementation, thereby limiting the scale and efficiency of the method in practical applications.