Aiming to address issues such as low security, inefficient encryption and decryption, weak robustness, image tampering, and the misuse of existing image encryption algorithms in blockchain-based applications for image sharing and transmission, this paper proposes a traceable image privacy protection scheme based on blockchain and homomorphic algorithm. Firstly, Paillier algorithm and image block technology were used to preprocess and encrypt the image. Feature images are extracted from encrypted images using lifting wavelet transform (LWT) and singular value decomposition (SVD). By means of dynamic S-box, binary watermark information is subjected to confusion, diffusion and XOR operations to generate zero-watermarking images for lossless embedding of traceability information. Then, the zero-watermarking image and its corresponding key are securely stored on the International File System (IPFS) through a smart contract mechanism to regulate image transactions and enable traceability of the image. Finally, the smart contract is utilized to detect any inaccuracies and errors in the zero-watermarking image, enabling tamper detection of said image. In experimental tests, the proposed scheme was shown to enable data distribution and computation with private data in ciphertext state, exhibiting high efficiency in encryption and decryption. Additionally, the zero-watermarking algorithm exhibits strong anti-attack capabilities and robustness. By considering image privacy protection and secure storage, tamper detection and traceability of encrypted images are effectively addressed.