Juvenile idiopathic arthritis (JIA) is the most common chronic arthropathy of childhood which is believed to be influenced by both genetic and environmental factors. The progress in identifying genes underlying JIA susceptibility using candidate gene association studies has been slow. Several associations between JIA and variants in the genes encoding the human leukocyte antigens (HLA) have been confirmed and replicated in independent cohorts. However it is clear that genetic variants outside the HLA also influence susceptibility to JIA. While a large number of non-HLA candidate genes have been tested for associations, only a handful of reported associations such as PTPN22 have been validated. In this review we discuss the principles behind genetic studies of complex traits like JIA, and comprehensively catalogue non-HLA candidate-gene association studies performed in JIA to date and review several validated associations. Most candidate gene studies are underpowered and do not detect associations, and those that do are often not replicated. We also discuss the principles behind genome-wide association studies and discuss possible implications for identifying genes underlying JIA. Finally we discuss several genetic variants underlying multiple clinically distinct autoimmune phenotypes.