The present study aimed to explore the relationship of obsessive-compulsive(OC) tendencies, confidence regarding inner experience and lack of feeling of knowing. We employed a 2(OC tendencies: high vs. low)×2(confidence: undermined vs. controlled) ×2(external proxies: upward phase vs. downward phase) mixed design, in which OC tendencies and confidence as the between-subjects variables, external proxies as the within-subjects variable. Padua Inventory(PI, Chinese version) was used to select 17 high(upper 10%,9 female and 8 male) and 18 low(lower 10%,9 female and 9 male) OC individuals. Then, we used instructions experimentally undermining/controlling confidence in judgments of the target internal state in non-selected participants of two groups. False biofeedback paradigm was used, in which relaxation as the target internal state, and two false pre-programmed biofeedback phases as external proxies for the internal state of relaxation, one of a descending line graph signaling to participants an increase in relaxation, and one of an ascending line graph signaling a decrease in relaxation. The order of the two false pre-programmed biofeedback phases was counterbalanced across participants and within each group. Participants were tested individually in a small and quiet room, we explored their situations of reliance on external proxies for internal states in different confidence conditions. Subjective assessments of relaxation were measured with a Visual Analogue Scale(VAS) on which participants were asked to place a mark that best described their relaxation levels during each task. Additionally, we measured their actual autonomic arousal level of galvanic skin response (GSR)and respiration(RSP) rate by a biofeedback apparatus(NeXus-10 Mark II, Mind Media B.V. Netherlands),in order to rule out a real relaxation level differences between the two biofeedback phases. At the end of the experiment, participants were asked to rate how confident they were about their two subjective relaxation estimates, using a scale of 0-100, in order to check whether our manipulation created a significant difference between participants with undermined/controlled confidence regarding their confidence in their estimates of internal states. The results showed that, (1)no matter how confident in their estimates of internal states, the high OC participants were more susceptible to external proxies in judging their internal states in comparison to low OC groups;(2)the high OC participants with undermined confidence were more affected by external proxies when judging their own internal states than with controlled confidence; in the condition of undermined confidence, high OC participants were more affected by external proxies in judging their internal states as compared to low OC participants. These findings indicated that,(1)individuals in high OC tendencies suffered reduced feeling of knowing in internal states;(2)high OC participants’ reduced feeling of knowing were affected by confidence regarding their inner experience. The research results can help to reduce high OC individuals’ doubt and related OC symptoms, and may be facilitate cognitive and meta-cognitive therapy of obsessive-compulsive disorder.