Background <p>Structured physical activity courses can effectively enhance preschool children’s motor skill development, and the type of activity may influence intervention outcomes. However, few studies have compared different courses, leaving educators without clear guidance for selecting effective interventions. This study examined the effects of three 12-week structured physical activity courses on the gross motor development of children aged 3–6, providing empirical evidence for intervention strategies.</p> Methods <p>A three-group quasi-experimental pretest–posttest design was adopted. Forty-five children aged 3–6 were recruited and assigned to one of three groups according to the course type: roller skating (RS), basketball (BT), and physical training (PT), with 15 children in each group. All participants received a 12-week intervention consisting of the assigned structured physical activity course, delivered twice weekly. Gross motor development was assessed before and after the intervention using the Test of Gross Motor Development-3 (TGMD-3). A 2 (time: pre vs. post intervention) × 3 (group: RS, BT, PT) repeated-measures ANOVA was used to examine within-group and between-group differences in TGMD-3 scores.</p> Results <p>After 12 weeks, the total gross motor score showed a significant time × group interaction. All three groups improved significantly (<i>p</i> &lt; 0.01), and post-intervention scores in the BT and PT groups were higher than in the RS group (<i>p</i> &lt; 0.01). The total locomotor score showed a significant main effect of time (<i>p</i> &lt; 0.01), with improvements observed across all groups. Significant improvements were observed across locomotor skills, although gallop and skip improved only in the BT and PT groups. The total object control score also showed a significant time × group interaction, with improvements in all groups (<i>p</i> &lt; 0.01) and higher post-intervention scores in the BT and PT groups compared with the RS group (<i>p</i> &lt; 0.01). Among individual object control skills, two-hand strike of a stationary ball, two hand catch, and underhand throw showed significant interactions, with BT and PT outperforming RS after the intervention (<i>p</i> &lt; 0.05). For the remaining skills, significant time main effects were detected across groups (<i>p</i> &lt; 0.01). All groups improved in one hand stationary dribble, forehand strike of a self-bounced ball, and kicking a stationary ball (<i>p</i> &lt; 0.01), while only the BT and PT groups improved in overhand throw (<i>p</i> &lt; 0.01). A group main effect was also observed for the one-hand stationary dribble test (<i>p</i> &lt; 0.01).</p> Conclusions <p>A 12-week engagement in roller skating, basketball, or physical training courses was associated with enhancements in locomotor skills, object control skills, and overall gross motor skill levels among 3–6-year-old children. Further comparison indicated that, although no significant group differences were observed in the overall gain in total locomotor scores, the basketball and physical training groups exhibited positive changes across all individual locomotor tasks. In contrast, while the roller skating group was associated with positive changes in most locomotor skills, less pronounced gains were observed in gallop and skip. Regarding object control skills, children in the basketball and physical training groups showed greater gains than those in the roller skating group in both overall object control performance and specific tasks such as two-hand strike of a stationary ball, two-hand catch, and underhand throw.</p>