This research compared three reinforcement learning (RL) algorithms (SAC, PPO, DDPG) to traditional PID control for water level control in single-tank and quadruple-tank systems. The RL algorithms were trained using MATLAB's Reinforcement Learning Toolbox and tested in a Simulink simulation.