Largely conducted in Western developed nations, research on community health screening has mainly been of limited duration. This study aims to ascertain the predictors of retention in a community health screening program, involving multiple admission cohorts over a 9-year period in Taiwan. Retention is defined as the participation in subsequent waves of health screening after being recruited for an initial screening. Data came from a prospective cohort study, named “Landseed Integrated Outreaching Neighborhood Screening (LIONS)”, in Taiwan. This research retrieved 5901 community-dwelling Taiwanese adults aged 30 and over from LIONS and examined their retention in three follow-ups during 2006–2014. Generalized estimating equations were employed to evaluate retention over time as a function of social determinants, health behaviors, and health conditions. Being middle-aged, higher education, and regular exercise were positively associated with retention. Conversely, smoking, betel-nut chewing, psychiatric disorder, hypertension, type 2 diabetes mellitus, stroke, and a longer time interval since enrollment were negatively associated with retention. Furthermore, retention rates varied substantially across admission cohorts with more recent cohorts having a lower rate of retention (aOR = 0.33–0.83). Greater attention needs to be directed to retention over time and variations across admission cohorts. Additionally, those who are in either younger or older age groups and have chronic diseases or unhealthy behaviors should be targeted with greater efforts.