This research is to study the solution to the defects in maintenance and inspection that can be predicted/prevented in advance among human factors that account for more than 70% of the causes of aviation accidents. Traditionally, mechanics have performed visual inspections of aircraft exteriors. Due to this, there were factors that affect the human ability of mechanics in aircraft maintenance and inspection, safety problems when performing the upper part of the aircraft inspection, and the difficulty of precise inspection. To improve these problems, we conduct a study on an AI drone inspection system that has deep-learned samples on aircraft damage/defects. In this paper, we describe the aircraft maintenance inspection checklist, non-destructive inspection types, types of aircraft damage and defects, deep-learning highly reliable AI drone inspection systems, and the expected effects of this technology and future applications. Through this system research, it is expected that mechanics will efficiently inspect the aircraft through the optimization of aircraft maintenance system technology to prevent aviation accidents in advance and reduce time and economic costs.