Health is influenced by the composition of diet as well as the total energy intake, and poor dietary choices, like excessive consumption of fat and carbohydrates, have been shown to cause health complications. Just as there are mechanisms to promote carbohydrate intake, there likely exists unknown pathways that reduce carbohydrate intake, and such interoceptive pathways have the potential to influence the quality of our food choices. In 2013, two genome-wide association studies (GWAS) both identified SNPs in the FGF21 locus associated with increased carbohydrate and decreased protein and fat consumption (Chu, 2013; Tanaka, 2013). This suggested a role for fibroblast growth factor 21 (FGF21) as a modulator of macronutrient intake. FGF21 is a peptide hormone, which mainly, but not exclusively, is produced by the liver, an organ which is uniquely positioned to sense circulating levels of nutrients and excrete signalling molecules to change ingestive behaviour. We, therefore, hypothesized that FGF21 might affect food preferences in mouse models. We increased circulating levels of FGF21 by either injecting or infusing human recombinant FGF21 or using a transgenic mouse model that overexpresses endogenous FGF21. We also developed whole-body and liver-specific FGF21 knockout mice. We made use of different choice paradigms to assess sugar preference. Mice were either given a choice between chow and a high-sucrose diet (HSD); or subjected to a two-bottle choice between water and a nutrient solution, for example, fructose. When endogenous FGF21 is knocked out, we observe a 2-fold increase in consumption of sugars (Suc, Fru, and Glu). Even the liver-specific FGF21 KO strain displays strong preference for sucrose-enriched diet and water, demonstrating that liver-derived FGF21 is responsible for regulating sugar intake. When increasing circulating levels of FGF21 levels, we observe decreased consumption of sugars. A preference towards other macronutrient classes, like fat and protein, was not found in two-bottle choice studies. Furthermore, no preference was found for complex carbohydrates, which warrants a specificity of FGF21 regulation to sugars and sweet-tasting compounds. Importantly, upon ingestion of sugars, we observe a physiological induction of FGF21 transcript and protein levels in WT mice. The FGF21 co-receptor, beta-klotho (KLB), has previously been shown to be expressed in different brain regions. To examine which brain regions FGF21 is signalling through to regulate sugar intake, KLB was knocked out in either the PVN, SCN or NTS. The FGF21-mediated reduction in sugar intake was attenuated only when KBL was deleted in the PVN. In addition, high plasma levels of FGF21 resulted in increased immunoreactivity of c-fos in the PVN, confirming the involvement of the hypothalamus in sucrose-specific appetite regulation. Due to the induction of circulating FGF21 by sugar intake, we propose a negative feedback loop, where the liver senses the amount of consumed sugar and down-regulates further intake through expression of FGF21. Knockout of hepatic ChREBP abolishes the induction of FGF21 transcription, suggesting that sugar activate the transcription factor ChREBP to stimulate FGF21 production. This study raises the possibility that sugar intake could be manipulated through molecular therapies, which could have promise in combating obesity and diabetes by enforcing the nutritive quality of dietary choices.