Cobalt (Co) is an essential element in humans as a component of vitamin B12. However, at high levels Co exposure has been shown to have detrimental effects. This study was designed to identify a chronic oral reference dose (RfD) for Co. Currently available data indicate that non-cancer health effects associated with Co exposure may include hematological, neurological, immunological, reproductive, cardiovascular, and endocrine responses. This analysis employs the standard US EPA risk assessment methodology for establishing a chronic RfD. In this analysis, the Jaimet and Thode (1955) 10-week, multiple dose human study of thyroid effects (decreased iodine uptake) in children was determined to be the most robust and sensitive study for identifying a potential point of departure dose (POD). A dose of 0.9 mg Co/kg-day was chosen as the POD. Consistent with the US EPA’s previous derivation of the perchlorate RfD, which is also based on decreased iodine uptake in humans, we considered several uncertainly factors (UFs), and determined that a factor of 10 for human variability was appropriate, as well as a factor of three for database adequacy. Applying an aggregate uncertainty factor of 30 to the POD yields a chronic oral RfD of 0.03 mg/kg-day. We believe this value would be protective of non-cancer health effects in the general population for a lifetime of daily exposure to Co.