: In today's modern world, the Internet is highly crucial. A larger number of wired and wireless sensors characterize the Internet of Things (IoT). These sensors are networked devices that create massive amounts of data, reaction time, latency, and security vulnerabilities. These problems create delay in decision making in some situation which is very extreme in case of Healthcare, Smart vehicles, Industry 4.0 etc., the data moving from the sensor devices have to reach the cloud and from there it has to reach back to receivers. Therefore, in order deal with such huge data we use fog computing, a well-known distributed architecture. Fog computing works with the aim to enhance the processing, intelligence, and accumulation of data closer to the Edge devices. The proposed framework helps in reduce latency as we place a Fog node device between the cloud and the edge device where data is generated and sent to the cloud and retrieved from the cloud. The framework is designed in such way that the different sensor devices can be placed on it and collects the sensor data from them. To test the framework functioning, without the fognode and with the fognode, comparing the latency with packet transfer rate from sensor devices to the cloud and vice versa. In this paper, we are considering two different case studies to test the proposed functioning of the framework with the following case studies i) Related to medical are like diabetes and cardiovascular illnesses are considered for prediction based on patient health records and ii) Vehicle theft where the owner get an alert of the theft of the vehicle within few seconds of the theft. In Medical case study initially, patient health data is acquired and stored using Sensor devices. A group of patient health records is first subjected to the unique rule-based clustering technique. Finally, we consider diabetes and cardiovascular diseases in our study. To test the performance of the suggested task, a rigorous experiment and study using healthcare data were conducted. Various Machine Learning (ML) algorithms are applied to patient health information, and the prediction result, as well as the accuracy of each algorithm, is generated. The findings show that the proposed work accurately predicts cardiac and diabetes problems, identifies which algorithm has the highest accuracy among them, and reduces burden on cloud. Devices having inbuilt sensors used in this work are pulse oximeter, smart watch, glucometer and thermometer. In Vehicle theft case study, there are numerous anti-theft solutions for smart vehicles based on GSM and cloud computing. The previous models have excessive latency and bandwidth. Fog computing is a creative new concept that combines existing solutions with GPS and GSM. IoT development for a fog- based vehicle anti-theft technology that uses a mobile application to pinpoint the exact position of a stolen vehicle. Low latency, low bandwidth, fast responsiveness, and high security are all advantages of fog computing. Thus with the use of the above proposed framework can reduce the latency and bandwidth in these applications and helps to avoid the delay decision making. Further this framework can be used to analyze and improve the performance and consumption of power in smart cities, smart homes and smart appliances..