The objective of this research is to identify the underlying components of English language learners' anxiety in online learning during the COVID-19 and to verify the coherence of the component model with the empirical data. A total of 408 Thai EFL university students, who were selected via a simple random sampling method, were assessed with the Online World Languages Anxiety Scale (OWLAS). An exploratory factor analysis (EFA) and a confirmatory factor analysis (CFA) were conducted using LISREL 8.80 to verify the factor structure of Online English Language Anxiety (OELA). The results of the EFA indicate six emerged underlying components: 1) use of English skills online, 2) negative feelings in online learning, 3) positive feelings in online learning, 4) online interaction, 5) cross-linguistic interference, and 6) attitudes to self and English classes. Considering the indices, the model was fit with the empirical data, with X[superscript 2] = 1115.24, df = 625, p-value = 0.048, X[superscript 2]/df = 1.784, GFI = 0.96, AGFI = 0.97, CFI = 0.99, SRMR = 0.043, RMSEA = 0.045. Factor loading values of each index were between 0.50-0.88. The indicators that had the highest weight value or most significantly influenced students' OELA were 'attitudes to self and English classes' and the least was 'cross-linguistic interference'.