Blockchain-based Land Records Verification System with Multiple Sales Fraud Detection using Decision Trees

The problem of land ownership fraud, especially the multiple sales of the same land parcel, has become a major challenge in property deals. Traditional land registration systems are often based on paper, centralized, and easy to tamper with, forge, or modify without approval. These outdated systems lack transparency, rely heavily on manual checks, and are open to insider fraud. This leads to disputes, legal issues, and financial loss for buyers. To tackle these problems, this paper suggests a blockchain-based framework for secure verification of land records, combined with a machine-learning fraud detection system. The blockchain provides decentralized, tamper-proof storage of land transactions by using smart contracts on the Ethereum network. Each land record is uniquely stored. Any change or duplication is impossible without agreement. In addition, the system uses a Decision Tree algorithm to spot fraudulent sales patterns based on transaction history, ownership legitimacy, and time-related features, allowing for early fraud detection before a transaction is finalized. The proposed solution makes land transactions secure and clear. Only verified owners can start a sale, and every transaction undergoes a fraud analysis before it is recorded on the blockchain. MetaMask handles user authentication and digital signatures, while Ganache offers a local Ethereum testing environment. Smart contracts manage land transfers, control access, and create unchangeable transaction logs. Thorough design analysis and simulated testing show that the system improves transparency, boosts fraud detection, and greatly reduces reliance on intermediaries. This research sets the groundwork for a decentralized, smart, and fraud-resistant land registry system that governments can adopt to ensure reliable and efficient property ownership verification in the digital age.