Mortality from breast cancer is principally due to tumor recurrence. Recurrent breast cancers arise from the pool of residual tumor cells, termed minimal residual disease, that survive treatment and may exist in a dormant state for up to 20 years following treatment of the primary tumor. As recurrent breast cancer is typically a fatal disease, understanding the mechanisms underlying dormant tumor cell survival is a critical priority in breast cancer research. Using a HER2/neu inducible bitransgenic mouse model, we demonstrate that the survival and recurrence of dormant mammary tumor cells is dependent on autophagy. We find that autophagy is induced in vivo following HER2/neu down-regulation and remains activated in dormant residual tumor cells. Using genetic and pharmacological approaches we demonstrate that inhibition of autophagy by chloroquine administration, Atg5 or Atg7 knockdown, or deletion of a single allele of the tumor suppressor Beclin 1 is sufficient to inhibit mammary tumor recurrence and that autophagy inhibition in dormant mammary tumor cells in vivo results in their death. Together, our findings establish a pro-tumorigenic role for autophagy in tumor dormancy and recurrence, demonstrate that dormant tumor cells are uniquely reliant upon autophagy for their survival, and reveal that targeting dormant residual tumor cells by inhibiting autophagy is sufficient to impair tumor recurrence. Additionally, we found that the mTOR inhibitor rapamycin delayed mammary tumor recurrence in this same mouse model by decreasing the survival of dormant mammary tumor cells. These studies identify pharmacological targets for a cellular state that is resistant to commonly used anti-neoplastic agents and suggest that autophagy inhibition as well as mTOR inhibition may be particularly effective in the context of dormant minimal residual disease.