Swarm intelligence can be described as a complex behaviour generated from a large number of individual agents, where each agent follows very simple rules. It is actually inspired by understanding the decentralized mechanisms in the organization of natural swarms such as the birds, the ants, the bees, the glowworms, and the fireflies. Observation of these biological behaviour has given birth to swarm robotics whereby robots have the capability to work with one another in a group to achieve the same kind of parallelism, robustness and collective capabilities. A collective behaviour movement strategy such as a “source search” and “aggregation” are commonly exhibited by the animals while finding their source of food. However, the situation for the robots is to find the source of odour, light, and sound. Meanwhile, there has been mounting interest, particularly for finding the deepest location in lakes and dams for bathymetric survey systems. Using the existing lawnmower methods incur substantial costs in terms of time, accuracy and reliability. Therefore, the usage of a swarming robotic system is proposed. In this thesis, a simple framework and methodology in developing a bio-inspired algorithm for cooperative swarming robotic application has been developed. The fruit flies or Drosophila Melanogaster movement strategy offers some advantages such as strategic 'search-aggregation' cycle, distribution of moving patterns with Levy Random, information sharing in real-time, and reduction of controller parameters during movements. A number of benchmark function processes were conducted to assess the performance of proposed FOA (Fly Optimisation Algorithm).