Among the various renewable energies, Solar energy is one of the most important sources. Solar panel is used to collect the solar energy and convert it into useful electrical energy. The dust accumulated on the solar panel reduce its efficiency to a certain degree. To overcome this problem, efficient techniques to clean the solar panel must be implemented. The proposed model is to clean the dust and bird droppings that has accumulated on the solar panel. An AI-based solar panel cleaning robot is designed which performs dry or wet cleaning based on the dirt or bird droppings, thus reducing water usage. The robot utilizes a Convolutional Neural Network model based on the Visual Geometry Group 16 architecture to detect dirt and bird droppings on solar panels. The robot is designed to perform dry cleaning using a brush and wet cleaning using a water pump, depending on the type of dirt detected. The experimental readings show that the power output of the solar panel increases significantly after cleaning, and the prototype model demonstrates the effectiveness of the robot in detecting and cleaning dirt and bird droppings. The development of such a robot has the potential to improve the efficiency of solar panels and reduce water usage in the cleaning process, making it a more sustainable and eco-friendly solution for maintaining solar panels.