The financial supply chain is vital for economic growth, enabling business expansion and innovation. However, it’s challenged by issues like lack of standardization, fraud risks, and limited transparency. The Privacy Preserving Blockchain Misbehavior Detection (PPBCMD) scheme aims to tackle these through blockchain technology for secure, transparent data sharing and storage. Cryptographic techniques like Zero Knowledge Proofs (ZKP) proves the authenticity of data without disclosing additional information, and Partial Homomorphic Encryption (PHE) to ensure confidentiality and perform computations on encrypted data provide a fine-grained approach to data security, ensuring that sensitive financial data remains secure and confidential. By employing privacy-trained ML models that utilize differential privacy using laplace noise on encrypted data, the proposed system ensures the confidentiality of sensitive fields while effectively detecting any instances of misbehavior within the financial supply chain characterized by unusually large transactions and those happening at atypical times. The system was validated using the PaySim dataset, with the Random Forest algorithm achieving a 97.59% F1 score after noise while preserving the sensitive financial data fields.