Although previous studies examined the association between mobility and disability, they have used either subjective measure disability such as activity of daily living or instrumental activity of daily living or indirect measure such as long-term care service use with small size of participants. This study aimed to examine the association between timed up and go (TUG) test and disability incidence with national disability registration data in Korea longitudinally, by using a national representative sample. We used the National Health Insurance Service-National Health Screening Cohort (NHIS-HEALS) database of National Health Information Database. The NHIS-HEALS dataset includes disability information of National Screening Programme participants, including registration date and type of disability, which is merged from Korean National Disability Registry (KNDR). We used Cox proportional hazard models to evaluate the association between TUG and disability incidence. We constructed three models with different levels of adjustment; Model 3 was a fully adjusted model. We conducted subgroup analysis according to the risk factors for disability. The study population comprised 81,473 participants; 86 of them were newly registered to KNDR, which were observed during a mean follow-up of 4.1 ± 2.6 (maximum, 8.9) years. For 334,200.9 person-year (PY) follow-up, the disability incidence rate was 0.208 per 1,000 PY. Disability incidence was significantly higher in participants with abnormal TUG results than in those with normal TUG results. (adjusted hazard ratio [aHR] 1.600, 95% confidence interval [CI] 1.036-2.472). In subgroup analysis, the disability incidence increased in participants of normal cognition, without obesity or without cardiovascular (CV) disease. Increased incidence in disability was noted in participants with abnormal TUG results. The increase was more evident for participants with normal cognition, without obesity or CV disease.