Planar underactuated mechanical systems are unaffected by gravity and therefore widely used in aerospace robots and underwater vehicles. However, the stable control is quite difficult since any position that the system can reach is its equilibrium point where it is linearly uncontrollable. In this paper, considering a planar three-link passive-active-active underactuated manipulator, a novel control approach using intelligent optimization is developed for its asymptotic stability at a target position. In the first stage, we develop controllers for stabilizing one active link at its target angle obtained by inverse kinematic method and the other two links at their transitional angles obtained by the Genetic Algorithm (GA), where the angles need to satisfy the angle constraint. Meanwhile, the GA is also utilized to optimize controllers’ parameters to ensure this stable control. Then in next stage, under the control of the angle constraint, the remaining two links are moved to the target angles simultaneously. Besides, we also improve the optimization criterion to include the control error, control time and input torque, thereby selecting better parameters of the controllers to obtain better control performance. Four simulations are implemented to prove the validity and feasibility of the proposed control and optimization algorithm.