Abstract: Online marketplaces have become integral parts of e-commerce, providing convenient platforms for buying and selling goods. Detecting fraudulent listings is crucial for maintaining trust and integrity within the marketplace. This project addresses this challenge by proposing a method for fraud detection based on predicting listing prices and identifying discrepancies betweenpredicted and listed prices and also updating prices based on current market conditions.By using machine learning techniques to predict listing prices based on working features,customer demand and market status, the aim is to uncover potentiallyfraudulent listings where the listed price significantly deviates from the predicted value. The project utilizes a dataset containinginformation about listings on the online marketplace, including features such as product category, description, and location, alongwith the listed prices. Data preprocessing techniques are applied to clean and prepare the dataset for analysis. Machine learning al-gorithms, including regression models and natural language pro- cessing techniques for textual data, are employed to predict listingprices based on their features.This also includes some methods offeature engineering. In conclusion, this project presents a novel approach to fraud detection in online marketplaces by leveragingmachine learning techniques for dynamic price prediction. By identifying discrepancies between predicted and listed prices, thedeveloped model effectively detects potentially fraudulent listings. The integration of this model into the marketplace platform enhances security and reliability, fostering a safer and more trustworthy environment for buyers and sellers. Moving forward,further refinement and optimization of the model can lead to evengreater accuracy and effectiveness in fraud detection.