With the development of a lot of delay-sensitive applications, it is very difficult to perform all the tasks locally because our devices have less computation power. Traditional Cloud Server (CS) has large computing power and storage but it is very far from IoT devices that’s why it consumes a lot of energy and take a huge time to execute a task. For these types of applications, local devices and CSs are not good choices. So, to overcome this delay-sensitive problem, the Fog computing (FC) concept was introduced. In FC, we can offload our task to Fog Nodes (FN) that are close to the user device and take less execution time and energy than a CS. To minimize the execution time for delay-sensitive applications, we have modeled our problem and converted our solutions from a bounded real space to a finite discrete space that uses the Flow Direction Algorithm (FDA). Our algorithm performance is 12 \(\text{\%}\) better than cloud-only, 6 \(\text{\%}\) better than Fog-only, and 6 \(\text{\%}\) better than IoT-only on average. In this, the task can execute locally or on a FN or CS. The proposed approach performed better to reduce the execution time.