As the demand for cross-departmental data collaboration continues to grow, traditional encryption methods struggle to balance data privacy with computational efficiency. This paper proposes a cross-departmental privacy-preserving computation framework based on BFV homomorphic encryption, threshold decryption, and blockchain technology. The proposed scheme leverages homomorphic encryption to enable secure computations between sales, finance, and taxation departments, ensuring that sensitive data remains encrypted throughout the entire process. A threshold decryption mechanism is employed to prevent single-point data leakage, while blockchain and IPFS are integrated to ensure verifiability and tamper-proof storage of computation results. Experimental results demonstrate that with 5,000 sample data entries, the framework performs efficiently and is highly scalable in key stages such as sales encryption, cost calculation, and tax assessment, thereby validating its practical feasibility and security.