Objectives  The idea of using the main dimensional factors of a carrier' hull to demonstrate a flight deck's primary design parameters and mathematical models was proposed.

  Methods   After analyzing the design data of thirteen foreign large-middle sized aircraft carriers by virtue of the support vector machine regression (SVR) method and the RBF kernel function, the regression analysis model for predicting the flight deck's primary design parameters was established, and the calculation arguments optimized by the simulated annealing (SA) algorithm to improve the precision of the regression. To examine the performance of the prediction model, the USS Kennedy (CV-67) was selected as the checking sample and the other twelve aircraft carriers as the training samples.

  Results  By comparing the prediction model's regression results with the actual design parameters of the USS Kennedy, it was found that the majority of the flight deck's primary design parameters were excellent and precise predictions. On the other hand, the reasons for the few imprecise predictions were provided in detail.

  Conclusions  The work in this paper provides the necessary technical support for the preliminary design of the aircraft carrier's flight deck.