Recent advances support the concept of autoimmune pancreatitis (AIP) as a unique systemic disease, because it shows occasional extrapancreatic lesions such as sclerosing cholangitis, sclerosing sialoadenitis, and retroperitoneal fibrosis, pathological features similar to those of fibrosis, and abundant infiltration of IgG4-positive plasma cells, and it is steroid responsive. Based on these findings, several diagnostic criteria have been proposed. Although AIP is accepted worldwide as a unique clinical entity, its pathogenetic mechanism remains unclear. To clarify its pathogenesis, its genetic background, humoral immunity, candidate target antigens including self-antigens and molecular mimicry by microbes, and cellular immunity including regulatory T cells, the complement system, and experimental models are reviewed. On the basis of this review, we hypothesize that the pathogenesis of AIP involves a biphasic mechanism consisting of “induction” and “progression.” In the early stage, the initial response to self-antigens [lactoferrin, carbonic anhydrase (CA)-II, CA-IV, pancreatic secretory trypsin inhibitor, and α-fodrin] and molecular mimicry (Helicobacter pylori) are induced by decreased naïve regulatory T cells (Tregs), and T-helper (Th) 1 cells release proinflammatory cytokines [interferon-γ, interleukin (IL)-1β, IL-2, and tumor necrosis factor α]. In the chronic stage, progression is supported by increased memory Tregs and Th2 immune responses. The classical complement system pathway may be activated by the IgG1 immune complex. As Tregs seem to play an important role in progression as well as in induction of the disease, further studies are necessary to clarify the pathogenesis of AIP.