This project focuses on implementing an adaptive filter using the Least Mean Squares (LMS) algorithm to remove echo from speech signals. The methodology involved: Generating an artificial echo by introducing a delay and attenuation to the original speech signal. Applying an LMS adaptive filter to estimate and subtract the echo. Producing a cleaned speech signal with reduced echo. Results: The LMS algorithm effectively minimized the echo, improving speech clarity. Frequency spectrum analysis confirmed the reduction of unwanted echo components. Future Enhancements: Implementing real-time echo cancellation for live applications. Exploring advanced filtering techniques such as Recursive Least Squares (RLS). Enhancing noise robustness for better performance in complex environments. This project demonstrates the effectiveness of adaptive filtering techniques in speech signal processing, with applications in telecommunications, video conferencing, and AI-based voice systems.