The “Double Reduction” policy mandates that compulsory education homework design achieve “burden reduction and quality improvement.” However, the “one-size-fits-all” approach remains prevalent in primary school English homework, failing to accommodate individual student differences. This study adopted a two-cycle action research paradigm, involving two fourth-grade classes from Primary School C in City P (50 students in the experimental class, 52 in the control class). Employing a quasi-experimental design with longitudinal tracking (pre-test, mid-test, post-test) within the experimental class and cross-sectional comparisons between the experimental and control classes, the study quantitatively examined the effects of differentiated homework design on student academic achievement. The results showed that the experimental class’s post-test average score significantly improved by 5.86 points compared to the pre-test (t = 23.53, p < 0.001), and the difference with the control class’s post-test score was also statistically significant (t = 2.58, p < 0.05). This effect accumulated progressively over the practice period. Furthermore, students in the experimental class reported a 14% increase in homework enjoyment and a 12% decrease in perceived burden, while the control class experienced a 13.5% increase in perceived burden. The study concludes that the differentiated homework design, through its “implicit differentiation + independent choice” implementation strategy and “1+X” flexible framework, can effectively enhance students’ academic performance and improve their attitudes towards homework, providing a reference for optimizing and empirically evaluating primary school English homework design under the “Double Reduction” policy.