The conventional anti-icing aeration vehicle is normally fixed to a certain position, which can only cover a small field when in operation. In view of this, a novel floating aeration vehicle that possesses both mobility and a wide operation coverage is proposed; however, research on its heating effect and operational optimization remains relatively insufficient. In order to improve the anti-icing performance of the floating aeration vehicle, an intelligent prediction model and a multi-objective optimization model were established for the aeration heating effect under various operating modes based on neural networks and multi-objective optimization algorithms. Schemes for various working conditions were derived by the models, including heating intensity biased scheme, balanced scheme, and endurance biased scheme. In comparison to the control group, the optimized schemes exhibited varying degrees of increase or decrease in the optimization of the anti-icing aeration vehicle. Among them, the heating effect of the heating intensity biased scheme increased by 79.12% compared with the control group, but the endurance time of the scheme decreased by 81.15%; the balanced scheme showed a minor improvement in both heating effect and endurance time; and the endurance time of the endurance biased scheme increased by 97.35%. The research findings can provide a technical support for the actual operation of anti-icing aeration vehicles.