In response to the lack of dynamic grading capability in traditional encryption mechanisms and the problem of single point failure and tampering diffusion in centralized storage, this study proposes secure storage and tamper proof technology for smart grid smart meter operation data. Based on weighted composite functions and information entropy algorithms, dynamically evaluate data sensitivity and classify it, and use elliptic curve cryptography (ECC) and advanced encryption standard (AES) algorithms to differentially encrypt data. Utilizing the Merkel Patricia tree (MPT) three-layer sharded storage structure to enhance security and tamper resistance. Generate digital fingerprints by combining Gaussian sequences with binary random codes, and accurately trace the source of tampering through similarity matching algorithms. The experimental results show that the proposed method can significantly improve the data write throughput, enhance the security of encrypted data, and achieve high-precision traceability even in the face of multiple tampering sources. The application effect is good.