Background: The World Health Organization has reinforced its recommendation to maintain preventive measures for COVID-19. The success of preventive measures is largely dependent on public attitudes and knowledge. Aims: This study aimed to assess the relationship between knowledge, attitude, risky behaviours and preventive measures and coronavirus 19 (COVID-19) infection in the Lebanese population. Methods: This cross-sectional study was conducted during September–October 2020 using the snowball sampling technique and an online self-administered questionnaire. The 4-part questionnaire targeted sociodemographic characteristics; medical history; knowledge, attitude and practices (preventive measures and risky behaviours related to COVID-19); and mental health variables such as psychological distress. Two models were derived using multivariable binomial logistic regression to optimize the picture of COVID-19 correlates. Results: Our sample comprised 1119 adult participants. Being older, female, a regular alcohol consumer, waterpipe smoker, having low level of education, low family income and having contact with a COVID-19 patient correlated with increased odds of ever having been diagnosed with COVID-19. Participants who had ever been diagnosed with COVID-19 had a significantly better current knowledge and a higher risky practice scale [adjusted odds ratio (ORa) = 1.49; 95% CI 1.27–1.74; P < 0.001; and ORa = 1.04; 95% CI 1.01–1.08; P = 0.024 respectively]. Conclusions: The most important predictors of COVID-19 infection appear to be generally well known, yet they should be continuously re-evaluated among the general population. This study highlights the need to raise proper awareness to improve precautionary behaviours among the public.