Background: Many prostate cancer survivors are treated with androgen deprivation therapy (ADT), but these therapies may increase frailty, worsen physical functioning, and increase fall risk. While exercise may counter functional declines associated with ADT, no studies have tested whether and which type of exercise may reduce falls and frailty. The purpose of this trial is to compare the relative efficacy of strength training versus tai ji quan training against each other and to a stretching control group on falls, frailty and physical functioning in men expose to ADT for prostate cancer. Methods: Prostate cancer survivors treated with ADT (n = 360) will be recruited to participate in this single-blind, parallel group, randomized trial. To be eligible men must meet the following criteria: 1) histologically confirmed prostate cancer, 2) received at least 6 months of ADT within the past 10 years, 3) report >1 fall in the past year OR have a score on one of two physical performance tests that is associated with increased fall risk (i.e. ≥ 12.0 seconds to complete the 3 meter timed up and go (TUG), or ≥ 10.0 seconds to complete 5 chair stands), 4) completed any other treatment at least 6 weeks prior to enrollment and not be on any concurrent prostate cancer therapy besides ADT, 5) not currently participating in moderate-vigorous intensity lower-body strength training or tai ji quan training ≥ 2 times/week for ≥ 30 minutes per session, 6) no cognitive difficulties that limit ability to answer survey questions or participate in exercise classes and performance tests, 7) no medical condition, disorder, or take medication that contraindicates participation in moderate intensity exercise, and 8) are able to communicate in English. Participants will be randomized to one of three supervised, group training programs: i) strength training, ii) tai ji quan training, or iii) stretching (control), that train 3x/week for 6 months. Outcomes are assessed at baseline, 3 (mid-intervention), 6 (immediately post-intervention) and 12 (follow-up) months. The primary outcome is falls assessed by monthly self-report. Secondary outcomes include the following: Frailty (low lean body mass (by bioelectrical impedance analysis), exhaustion (by SF-36 vitality scale), low activity (by CHAMPS physical activity survey), slowness (by 4m usual walk speed), and weakness (by chair stand time)); Objective and subjective measures of physical function will also be collected. Negative binomial regression models will be used to assess differences in falls between groups, while mixed-effects modeling will be used to compare the relative efficacy of training group on secondary outcomes. Registered November 18, 2018. Clinical trial information: NCT03741335.