Background: Neurocognitive deficits are noted in anxiety disorders (ADs), albeit with several inconsistencies. The relationship between neurocognition and metacognition may have potential implications for understanding cognitive dysfunction but is poorly understood in ADs. This study aimed to examine the relationship between neurocognition and metacognition in ADs, with a cross-sectional design. Methods: The sample included ADs ( n = 25) and nonclinical ( n = 25) groups matched on age, gender, and education. Neurocognition was assessed using tests for intelligence, attention, working memory, fluency, flexibility, set-shifting, inhibition, planning, and memory; and metacognition using Metacognition Questionnaire-30 (MCQ-30) and Metacognitive Awareness and Regulation Scale (MARS). Results: Compared to comparison/normative scores, the anxiety group showed significantly poorer performance on zoo map test (low demand trial; P = 0.007), rule shift cards 1 (P ≤ 0.001), rule shift cards 2 (P ≤ 0.001), and logical memory immediate recall (P ≤ 0.001) and delayed recall (P ≤ 0.001); greater negative beliefs about worry (P = 0.005), and poorer metacognitive awareness and regulation (P = 0.01). Greater cognitive self-consciousness was correlated with better planning (Spearman’s rho = −0.509, P = 0.009). Conclusions: Individuals with ADs show neurocognitive difficulties in planning, set-shifting, and logical memory, dysfunctional metacognition, and reduced metacognitive awareness and regulation. Cognitive self-consciousness is linked to better planning. The interrelationships between neurocognition and metacognition may have potential implications for clarifying inconsistent findings and designing novel cognitive interventions in ADs.