There is great interest in developing hyperspectral imaging (HSI) techniques for rapid and nondestructive inspection of food quality, safety, and authenticity. In recent years, image quality has been constantly improved through advances in instrumentation, particularly in more powerful detectors. Nevertheless, pretreatment of data by de-noising is a necessary step to insure clean HSI datasets for further analysis. This review first introduces the typical and commonly used de-noising methods in HSI that correct for undesirable variations and remove noisy variables. Their advantages, disadvantages, and implementation are also discussed by giving examples of recent applications in the food industry. Finally, some advice is given for selecting the de-noising methods that are best suited for a particular application. This review offers an overview of the most frequently applied methods and the latest progress made in HSI de-noising in food applications. It provides systematic insight into future trends for generating high-accuracy predictions regarding food safety and quality.