Objective: to identify the socio-demographic, social connectedness, physical environment and physical and mental health-related factors associated with PA. Setting and participants: adults aged 65 and over living in the community (n = 3499). Methods: this study used a cross-sectional design. Self-report, interview and physical assessment were the data collection methods used. The International Physical Activity Questionnaire (short form) measured PA in metabolic equivalents. Results: 31.8% of older people did not meet the recommended guidelines. The regression model was significant, explaining 31.3% of the variance in PA (F = 34.32, P < 0.001). In order of strength of relationship, the beta coefficients, the variables time spent sitting, gender, quality of life, grip test, disability, age, depression, BMI, anxiety, employment, member of a non-church club, type of house and cognitive ability were significant factors associated with PA. Conclusion: from a range of 40 original variables, this study identified the variables most strongly related to PA were not physical health or physical environment. They included modifiable variables, such as time spent sitting and mental health, which can be actively targeted in policy and practice. They also included non-modifiable variables such as gender and age that can be addressed through increased awareness and targeted health promotion.