Unfounded beliefs (UB; i.e. beliefs that are not supported by scientific evidence), such as paranormal and conspiracy beliefs remain prevalent globally. Far from being innocuous, UB pose various public health (e.g. vaccine hesitancy) and societal hazards (e.g. anti-democratic attitudes). Despite mounting concerns in the current context of growing misinformation and the so-called infodemic, there remains a lack of effective educational interventions suitable for K-12 contexts to prevent UB among children. In this paper, we report the results of a phase I randomized-controlled trial assessing a novel critical thinking (CT) education intervention among secondary school students in France (8th and 9th graders, n = 135). The intervention consisted in eight weekly, 1-hour long classes. Its content was designed to improve children’s capability, favourable attitudes, and motivation towards exerting rational judgment, to impact the cognitive processes at play behind UB (e.g., overreliance on intuition, ontological confusions, underestimation of randomness as a causal force). Relative to placebo (class as usual), children in the intervention group displayed less conspiracy beliefs immediately post-intervention (d = .55) and one-month afterwards (d = .58, follow-up). The intervention had globally a stronger immediate effect on paranormal beliefs levels, (d = .74), but these reverted to pre-intervention levels one month after. To the best of our knowledge, this study is the very first randomized trial to demonstrate the effectiveness of a feasible, standardized CT education intervention in decreasing both paranormal and conspiracy beliefs among secondary school students.