The present randomized controlled trial examines the effects of yoga, as opposed to general physical education (PE) and passive controls, on the motor skills and self-esteem of Tunisian kindergarten children. Fifty-four children (24 females, 30 males; age 5.2 ± 0.6 years; weight 17.8 ± 2.3 kg; height 103.1 ± 5.0 cm; BMI 17.26 ± 0.46 kg/m2) were randomly divided into three groups of 18 (yoga, PE, and controls), and blindly administered the Eurofit physical fitness test battery and the Rosenberg Self-Esteem Scale prior to and after 12 weeks. During this period, the yoga and PE groups attended two weekly yoga and PE sessions of 30 minutes each, respectively. At T0, no significant differences between the groups were observed (p > 0.660). Repeated-measures analysis of variance associated yoga with higher levels of motor skills (p < 0.05; effect size for arm strength [ηp2] = 0.17, leg strength [ηp2] = 0.17, flexibility [ηp2] = 0.33; speed [ηp2] = 0.61, and balance [ηp2] = 0.19) and global self-esteem in comparison to the PE and control groups (p < 0.001; ηp2 = 0.42). Twelve weeks of kindergarten-based yoga appears to improve self-esteem and motor skills in 5-year-old children. The yoga intervention as a school routine seems to enhance development, health, and the well-being of children.