In recent years, online transactions of digital collections have been increasing, with platforms such as Alibaba Auctions and OpenSea facilitating their circulation in the market. However, the bidder’s bidding privacy is at risk of being disclosed during an online auction. To address this issue, this study proposes a privacy-preserving online auction approach based on the homomorphic property of SM2, which not only protects the users’ bidding privacy but also ensures the usability of the bidding data. Specifically, this study creates a homomorphic encryption scheme based on SM2, encrypting bidders’ bidding information and constructing a piece of noisy bidding information to conceal the privacy data. The efficiency of the online auction privacy preservation approach is improved by integrating the Chinese reminder theorem and baby step giant step (CRT-BSGS) into the homomorphic encryption process with SM2, which has proved to be more efficient than the Paillier algorithm. Finally, the security and efficiency of the proposed scheme are verified in detail.