Autoantigen treatment has been tried for the prevention of type 1 diabetes (T1D) and to preserve residual beta-cell function in patients with a recent onset of the disease. In experimental animal models, efficacy was good, but was insufficient in human subjects. Besides the possible minor efficacy of peroral insulin in high-risk individuals to prevent T1D, autoantigen prevention trials have failed. Other studies on autoantigen prevention and intervention at diagnosis are ongoing. One problem is to select autoantigen/s; others are dose and route. Oral administration may be improved by using different vehicles. Proinsulin peptide therapy in patients with T1D has shown possible minor efficacy. In patients with newly diagnosed T1D, subcutaneous injection of glutamic acid decarboxylase (GAD) bound to alum hydroxide (GAD-alum) can likely preserve beta-cell function, but the therapeutic effect needs to be improved. Intra-lymphatic administration may be a better alternative than subcutaneous administration, and combination therapy might improve efficacy. This review elucidates some actual problems of autoantigen therapy in the prevention and/or early intervention of type 1 diabetes.