Cancer stem cells (CSCs) have been proposed to be responsible for tumor recurrence, distant metastasis and drug-resistance, in the vast majority of cancer patients. Therefore, there is an urgent need to identify new drugs that can target and eradicate CSCs. To identify new molecular targets that are unique to CSCs, we previously compared MCF7 2D-monolayers with 3D-mammospheres, which are enriched in CSCs. We observed that 25 mitochondrial-related proteins were >100-fold over-expressed in 3D-mammospheres. Here, we used these 25 proteins to derive short gene signatures to predict distant metastasis (in N=1,395 patients) and tumor recurrence (in N=3,082 patients), by employing a large collection of transcriptional profiling data from ER(+) breast cancer patients. This analysis resulted in a 4-gene signature for predicting distant metastasis, with a hazard ratio of 1.91-fold (P=2.2e-08). This provides clinical evidence to support a role for CSC mitochondria in metastatic dissemination. Next, we employed a panel of mitochondrial inhibitors, previously shown to target mitochondria and selectively inhibit 3D-mammosphere formation in MCF7 cells and cell migration in MDA-MB-231 cells. Remarkably, these five mitochondrial inhibitors had only minor effects or no effect on MDA-MB-231 tumor formation, but preferentially and selectively inhibited tumor cell metastasis, without causing significant toxicity. Mechanistically, all five mitochondrial inhibitors have been previously shown to induce ATP-depletion in cancer cells. Since 3 of these 5 inhibitors were designed to target the large mitochondrial ribosome, we next interrogated whether genes encoding the large mitochondrial ribosomal proteins (MRPL) also show prognostic value in the prediction of distant metastasis in both ER(+) and ER(-) breast cancer patients. Interestingly, gene signatures composed of 6 to 9 MRPL mRNA-transcripts were indeed sufficient to predict distant metastasis, tumor recurrence and Tamoxifen resistance. These gene signatures could be useful as companion diagnostics to assess which patients may benefit most from anti-mito-ribosome therapy. Overall, our studies provide the necessary proof-of-concept, and in vivo functional evidence, that mitochondrial inhibitors can successfully and selectively target the biological process of cancer cell metastasis. Ultimately, we envision that mitochondrial inhibitors could be employed to develop new treatment protocols, for clinically providing metastasis prophylaxis, to help prevent poor clinical outcomes in cancer patients.