This research investigates the effectiveness of live walls in reducing heat transfer rates compared to conventional walls, using both experimental methods and machine learning techniques. The study aims to assess how different wall types (red brick, cement, plaster of Paris), insulating materials (cork sheet, rock wool, wood wool), and plant species (Pothos, Philodendron, Boston fern) influence temperature and humidity. Utilizing Arduino Mega, DS18B20, and DHT11 sensors, temperature, and humidity data were collected for 48 combinations of wall, insulation, and plant types. The experimental setup involved monitoring inside and outside conditions to measure heat transfer rates. Data analysis revealed a significant temperature differential, with the best combination identified as a plaster of Paris wall with wood wool insulation and Boston fern, achieving the lowest inside temperature. A random forest regressor model was trained using the collected data, achieving a high accuracy with an R2 of 0.945, indicating strong predictive capabilities.