The androgen deprivation therapy (ADT) to systematically suppress/reduce androgens binding to the androgen receptor (AR) has been the standard therapy for prostate cancer (PCa); yet, most of ADT eventually fails leading to the recurrence of castration resistant PCa. Here, we found that the PCa patients who received ADT had increased PCa stem/progenitor cell population. The addition of the anti-androgen, Casodex, or AR-siRNA in various PCa cells led to increased stem/progenitor cells, whereas, in contrast, the addition of functional AR led to decreased stem/progenitor cell population but increased non-stem/progenitor cell population, suggesting that AR functions differentially in PCa stem/progenitor vs. non-stem/progenitor cells. Therefore, the current ADT might result in an undesired expansion of PCa stem/progenitor cell population, which explains why this therapy fails. Using various human PCa cell lines and three different mouse models, we concluded that targeting PCa non-stem/progenitor cells with AR degradation enhancer ASC-J9 and targeting PCa stem/progenitor cells with 5-azathioprine and γ-tocotrienol resulted in a significant suppression of the tumors at the castration resistant stage. This suggests that a combinational therapy that simultaneously targets both stem/progenitor and non-stem/progenitor cells will lead to better therapeutic efficacy and may become a new therapy to battle the PCa before and after castration resistant stages.