With the rapid development of cloud computing technology, more and more companies and individuals choose to store data in the cloud, but this also makes the security of cloud data an urgent issue to be addressed. Security threats such as data leakage, illegal access and tampering are becoming increasingly severe, seriously threatening the privacy rights of users and the asset security of enterprises. This paper combines multiple encryption algorithms such as symmetric encryption, asymmetric encryption and homomorphic encryption to build a rigorous multi-level encryption system. First, symmetric encryption is used to perform preliminary data processing to improve the efficiency of data processing; secondly, asymmetric encryption technology is used to protect the symmetric key to ensure the security of the key during transmission; finally, homomorphic encryption technology is introduced to enable specific computing operations on encrypted data, thereby ensuring data privacy while also supporting data analysis needs. Through simulation experiments and comparative analysis, it is fully verified that the data security of this paper can reach up to 1, and the encryption rate can reach up to 1000bps. The experimental results show that this paper can greatly improve the security of cloud data while still maintaining a high computing efficiency.