Objective: To evaluate the effectiveness of a group-based 6-week intervention based on the OTAGO Exercise Program, to improve strength, mobility, gait and balance parameters in community-dwelling older adults. Methodology: A randomized controlled clinical trial with 196 participants in the intervention group (IG) and 205 in the control group (CG), community-dwelling older adults 75 to 90 years old, recruited in Primary Care Centres, excluding dementia, blindness, amputees and/or lower limb surgery 6 months previous, active neoplasms, institutionalized or included in home care program, as well as those with contraindication of physical exercise. At baseline, at 3 and 12 months after the end of the intervention, grip strength was evaluated with a Jamar® dynamometer, power (P) of the lower extremities with a Chronojump® linear encoder, gait with the sensor BTS G-Walk® and the Timed Up-and-Go test (TUG), balance with the Biodex Balance System® platform, the Tinetti scale and the Berg balance scale. Simultaneously, a once every three months telephone monitoring of the falls and consequences was carried out. Results: At baseline, 401 participants were analysed with physical tests (55. 1% women, mean age 80. 7 years), 237 at the 3 months follow-up, 127 at 12-month follow up, and 393 participants were monitored for falls. Age was significantly correlated (p<0. 05) with all physical variables except for the Fall Risk test (FR) and the sensory integration balance test with eyes closed and unstable surface (m-CTSIB CE). Women obtained lower values than men in all physical tests. The intervention achieved significant improvements (p=0. 02) in balance at post-intervention in men, evaluated using the Tinetti scale. Male participants in the IG significantly improved (p<0. 05) balance evaluated with the FR at 3 months and the m-CTSIB CE at 12 months follow up, as well as grip strength, gait speed and stride length, although without statistical significance. Women in the IG significantly improved balance (p=0. 01) evaluated with the Tinetti scale and the FR at 3 months (p=0. 04) and the m-CTSIB CE and stride length at 12 months (p=0. 01) follow up. No improvements in P and TUG score were observed in both sexes. The rate of falls was significantly lower (p<0. 05) in the IG, with this decrease being significant in women and in the age category of 80-84 years. In the same group, the rate of fractures numbers was significantly lower (p<0. 05) in both sexes and in the 80-84 years-old category, as well as the use of rehabilitation assistance resource (p<0. 001) and fear of falling (p<0. 001). Conclusions: The exercise program significantly improves balance in men at 3 months and has a tendency to improve grip strength, gait speed, cadence and stride length without being significant. In women, this favourable trend occurs in balance, gait speed and stride length without being significant. At 12 months follow-up, in both sex, grip strength, gait and balance achieve less loss, not significant. Older age is significantly correlated with all physical variables except the FR and m-CTSIB CE. Men obtained significantly higher scores than women except in gait speed, cadence and some balance test. The intervention is also effective in reducing falls in women and in the 80-84 years-old category, as well as their consequences.