In this study, a speaking plant-based optimal control system was proposed and applied to promote the initial growth of tomatoes in hydroponics. The control system consisted of a feedback control system and a decision system. The decision system consisting of neural networks and genetic algorithms was used to determine the optimal l-step set point of nutrient concentration which maximizes the initial growth of tomatoes. In the decision system, the growth rate of plant height to nutrient concentration was first identified using neural networks and then the optimal 6-step set points were determined through simulating the identified model using genetic algorithms. One step is 7 days. The optimal value (6-step set points) was 1.0 for the 1st step, 0.5 for the 2nd step, 0.8 for the 3rd step, 0.9 for the 4th step, 1.1 for the 5th step, and 1.2 dS m⁻¹ for the 6th step during the initial growth stage. There was a significant reduction (0.5 dS m⁻¹) in nutrient concentration in the second step and this significant reduction corresponds to nutrient stress. Actual plant growth for optimal control was about 1.15 times larger than that for conventional control. We suggest that this control technique is suitable for optimizing hydroponic cultivation processes and the control strategy, including nutrient stress application, is effective in promoting plant growth.