To address the challenge of improving downrange prediction accuracy while reducing unnecessary bank reversals, we propose a predictor-corrector guidance method based on high-precision analytical solutions. First, precise analytical solutions for downrange and crossrange are derived to establish the foundation for the guidance law, and we use a neural network model for accurate flight-path angle prediction. Next, we compensate for Earth's rotation effects in the downrange prediction by employing the Taylor expansion. Aiming to eliminate phugoid oscillations, we add the flight-path angle feedback into the command of the bank angle and angle of attack. Subsequently, a lateral guidance law that combines the heading angle error corridor with predicting bank reversals is introduced. The sign of the bank angle is determined by the corridor during the initial flight phase. Following the vehicle travels a specified downrange distance, we use the crossrange analytical solution to regulate subsequent bank reversals. Simulations under disturbance conditions demonstrate that the proposed guidance method has high accuracy and strong robustness, effectively meeting various constraints.