Crop losses due to pests can devastate agriculture, with up to 30% of yields lost yearly. However, pest detection in real world, complex and uncontrolled environment poses a significant challenge. This research paper provides an in-depth analysis of methods and techniques applicable for detecting pests in fruits and vegetables through the use of computer vision (CNN and YOLO), sensor-based techniques, and advanced learning algorithms. Our examination showcases strategies in AI-based pest recognition that lay the groundwork for effective pest detection and classification methods and showcases a comparative analysis of their advantages, limitations, and their results. The authors also outline future direction since the goal of this research study is to guide future researchers in improving pest control measures and help in building strong and resilient pest detection systems.