The rapid development of artificial intelligence technology has led to the proliferation of computational thinking (CT) education. However, research on unplugged activities’ influence in elementary math classrooms is limited, despite some exploring programming's cognitive benefits. This study presents both qualitative and quantitative analyzes stemming from a ten-week quasi-experimental research endeavor, specifically tailored for third-grade students. The research devised a series of unplugged activities, encompassing mathematical games, hands-on construction of mathematical logic boards, and calculating shopping discounts. The overarching objective was to investigate the impact of these unplugged activities on students’ mathematical creativity and CT. Students participating in a mathematics curriculum based on unplugged activities (N = 47) were compared with students participating in a traditional lecture-based mathematics curriculum (N = 46). The results indicated that unplugged activities exhibited significant advantages in fostering students’ CT and mathematical creativity across three dimensions, namely, problem-posing creativity (t = 5.830, p < 0.01), problem-solving creativity (t = 6.633, p < 0.01), and creative self-efficacy (t = 7.554, p < 0.01). Furthermore, the study revealed a relationship between students’ mathematical creativity as a predictor of CT. The results of the quantitative analysis were supported by the teacher's and students’ interview data, and the students felt excited and interested in the unplugged activities, whereas some students using the lecture-based method reported boredom and lack of interactivity. This research offered valuable insights for mathematics and CT education practice, underscoring unplugged activity as an innovative instructional approach that brings forth new possibilities for traditional mathematics teaching, with potential applications in the K-12 curriculum.