The application provides an
attack angle control method using a neural network and a high-precision analytical solution, and comprises the following steps: step 1, a perturbation model is established; the
terminal guidance of a
missile is considered as a large
earth model, therefore, a dynamic model of the
missile is established and small quantities therein are analyzed, a perturbation equation is established, and is used for subsequent solution of a trajectory analytical solution; step 2, zero-order solution solving; after the perturbation equation is obtained, an analytical solution of a zero-order term in the perturbation equation is obtained by using a
perturbation method, approximate fitting and mathematical derivation, and is used for subsequent solution of a first-order solution; step 3, first-order solution solving; after the solution of the zero-order term is obtained, a new perturbation equation is further constructed by using the
perturbation method, and an analytical solution of the first-order term is solved, and is used for calculation of a guidance law; step 4, resistance coefficient fitting using a neural network; given a
simulation state, different resistance change values are obtained through Monte Carlo target
simulation, fitting is performed through a quadratic function, fitting coefficients are obtained, and the fitting coefficients are fitted through a neural network; and step 5,
attack angle control guidance law; after the above analytical solution formula is obtained, the guidance law needs to measure a current state of the
missile, predict a terminal flight deviation, finally, correction is performed through a bias proportional guide, and the
attack angle is constrained.