Flocking behaviour, a widespread phenomenon in the natural world, represents coordination and collective motion observed among diverse species. Traditional approaches are mostly used to model this behaviour. However, these approaches rely on static flocking rules, limiting their adaptability to dynamic real-world scenarios. The challenge lies in effectively understanding and using this complex behaviour for practical applications. In this study, we present an approach using reinforcement learning to address this challenge. Our aim is to train autonomous agents to replicate flocking behaviour within a continuous 2D environment. The approach involves using a reward function to imitate flocking behaviour with an artificially generated flock. By overcoming these limitations, our study offers a deeper understanding of natural systems and broadens the scope for controlling swarming behaviours in various domains and environments.