IMPORTANCE: Seroprevalence studies inform the extent of infection and assist evaluation of mitigation strategies for the COVID-19 pandemic. OBJECTIVE: To estimate the prevalence of unidentified SARS-CoV-2 infection in the general population of Hong Kong. DESIGN, SETTING, AND PARTICIPANTS: A prospective cross-sectional study was conducted in Hong Kong after each major wave of the COVID-19 pandemic (April 21 to July 7, 2020; September 29 to November 23, 2020; and January 15 to April 18, 2021). Adults (age ≥18 years) who had not been diagnosed with COVID-19 were recruited during each period, and their sociodemographic information, symptoms, travel, contact, quarantine, and COVID-19 testing history were collected. MAIN OUTCOMES AND MEASURES: The main outcome was prevalence of SARS-CoV-2 infection. SARS-CoV-2 IgG antibodies were detected by an enzyme-linked immunosorbent assay based on spike (S1/S2) protein, followed by confirmation with a commercial electrochemiluminescence immunoassay based on the receptor binding domain of spike protein. RESULTS: The study enrolled 4198 participants (2539 [60%] female; median age, 50 years [IQR, 25 years]), including 903 (22%), 1046 (25%), and 2249 (53%) during April 21 to July 7, 2020; during September 29 to November 23, 2020; and during January 15 to April 18, 2021, respectively. The numbers of participants aged 18 to 39 years, 40 to 59 years, and 60 years or older were 1328 (32%), 1645 (39%), and 1225 (29%), respectively. Among the participants, 2444 (58%) stayed in Hong Kong since November 2019 and 2094 (50%) had negative SARS-CoV-2 RNA test results. Only 170 (4%) reported ever having contact with individuals with confirmed cases, and 5% had been isolated or quarantined. Most (2803 [67%]) did not recall any illnesses, whereas 737 (18%), 212 (5%), and 385 (9%) had experienced respiratory symptoms, gastrointestinal symptoms, or both, respectively, before testing. Six participants were confirmed to be positive for anti-SARS-CoV-2 IgG; the adjusted prevalence of unidentified infection was 0.15% (95% CI, 0.06%-0.32%). Extrapolating these findings to the whole population, there were fewer than 1.9 unidentified infections for every recorded confirmed case. The overall prevalence of SARS-CoV-2 infection in Hong Kong before the roll out of vaccination was less than 0.45%. CONCLUSIONS AND RELEVANCE: In this cross-sectional study of participants from the general public in Hong Kong, the prevalence of unidentified SARS-CoV-2 infection was low after 3 major waves of the pandemic, suggesting the success of the pandemic mitigation by stringent isolation and quarantine policies even without complete city lockdown. More than 99.5% of the general population of Hong Kong remain naive to SARS-CoV-2, highlighting the urgent need to achieve high vaccine coverage.