In this paper, we have presented a smart eco-friendly cooking model by integrating IoT and AI-ML models. This method introduces the sensor-based collection of real-time data on energy, water, and gas usage associated with cooking by capturing a wide array of data in conjunction. The data is processed in a multi-tiered computing architecture (cloud, fog, and edge computing layers) to achieve efficient data management and fast response. The system learns about the preferences and specific cooking behaviors of a user over time thereby customizing itself to each user and maintaining its effectiveness and usability across a long period of time. The model also has resource optimization strategies that alert users when they use resource-consuming cooking methods without compromising quality and automatically adjust the appliance settings. Meanwhile, polynomial regression improves the accuracy of prediction as compared to other regression model due to high nonlinearity between cooking variable which can also increase overall improve performance of system.