The paper introduces a high-performance single-stage PV-RO desalination system that operates without a battery, reducing system complexity, cost, and maintenance requirements. The primary focus is designing and controlling a low-voltage squirrel cage asynchronous motor powered by renewable photovoltaic energy. The system utilizes the Horse Herd Optimization Algorithm (HOA) for MPPT to optimize the power balance between the motor and solar panels, ensuring operation at the GMPP. Additionally, the motor's rotational speed is managed using Predictive Torque Control (PTC), based on predicting the system's future actions over a prediction horizon using the system model. The proposed control scheme is simulated using MATLAB/Simulink software, evaluating the system's performance under uniform solar radiation and partial shading conditions. The simulation results demonstrate that the HOA outperforms other MPPT methods, including the GWO, FPA, CSA, and P&O, in speed, efficiency, and convergence. Additionally, the well-established benefits of PTC, including precise torque response, reduced torque ripples, and accurate speed response, are verified through simulation results. The seamless integration of a single-stage design, enhanced power extraction, and precise motor speed control positions the proposed system as a promising technical and economical solution for PV/RO desalination systems. At last, its implementation feasibility is validated by a hardware-in-the-loop (HIL) experiment based on the DSpace 1104 platform.