Purpose:　There　is　an　unmet　clinical　need　for　an　accurate　and　objective　diagnostic　tool　for　early　detection　of　acute　exacerbation　of　chronic　obstructive　pulmonary　disease　(AECOPD).　DETECT　(NCT03556475)　was　a　multicenter,　observational,　cross-sectional　study　aiming　to　develop　and　validate　multivariable　prediction　models　for　AECOPD　occurrence　and　severity　in　patients　with　chronic　obstructive　pulmonary　disease　(COPD)　in　China.　Patients　and　Methods:Patients　aged　＞=　40　years　with　moderate/severe　COPD,　AECOPD,　or　no　COPD　were　consecutively　enrolled　between　April　22,　2020,　and　January　18,　2021,　across　seven　study　sites　in　China.　Multivariable　prediction　models　were　constructed　to　identify　AECOPD　occurrence　(primary　outcome)　and　AECOPD　severity　(secondary　outcome).　Candidate　variables　were　selected　using　a　stepwise　procedure,　and　the　bootstrap　method　was　used　for　internal　model　validation.　Results:　Among　299　patients　enrolled,　246　were　included　in　the　final　analysis,　of　whom　30.1%,　40.7%,　and　29.3%　had　COPD,　AECOPD,　or　no　COPD,　respectively.　Mean　age　was　64.1　years.　Variables　significantly　associated　with　AECOPD　occurrence　(P＜0.05)　and　severity　(P＜0.05)　in　the　final　models　included　COPD　disease-related　characteristics,　as　well　as　signs　and　symptoms.　Based　on　cutoff　values　of　0.374　and　0.405　for　primary　and　secondary　models,　respectively,　the　performance　of　the　primary　model　constructed　to　identify　AECOPD　occurrence　(AUC:　0.86;　sensitivity:　0.84;　specificity:　0.77),　and　of　the　secondary　model　for　AECOPD　severity　(AUC:　0.81;　sensitivity:　0.90;　specificity:　0.73)　indicated　high　diagnostic　accuracy　and　clinical　applicability.Conclusion:　By　leveraging　easy-to-collect　patient　and　disease　data,　we　developed　identification　tools　that　can　be　used　for　timely　detection　of　AECOPD　and　its　severity.　These　tools　may　help　physicians　diagnose　AECOPD　in　a　timely　manner,　before　further　disease　progression　and　possible　hospitalizations.