The assessment of food authenticity is a topic that gained a lot of interest at the international level. This term includes misidentification of variety, origin, production system, processing but also adulteration. These frauds all have an analytical component, and research tends to offer new analytical solutions to manage them. One of them is non-targeted approaches, which get around the limitations of targeted analysis by detecting the unexpected. A wide range of products are studied such as wine, rice, olive oil, spices, and honey among the top five. Geographic origin is by far the fraud with the most attention. The main reason is probably the complexity to consider terroir effect and every other variable to determine an area of production. This review offers an overview of the potential of non-targeted analysis to assess food authenticity. These results also illustrate the capability to look for environmental terroir markers that could be cross-matrixes.