This study aimed to identify early critical points of accelerated hyperopia reserve depletion in Chinese children with poor visual habits using Dynamic Network Landscape Entropy (DNLE). We hypothesized that DNLE could detect critical transitions in ocular parameter networks before myopia onset, enabling timely interventions.A prospective cohort study (Anyang Childhood Eye Study, 2012-2019) with 6 annual follow-ups. Participants were categorized into 3 groups based on risk factors: prolonged near work, insufficient sleep, or limited outdoor activity.About 3112 grade 1 students were screened, leaving 2855 eligible participants (57.2% female). After selecting 3 groups with poor visual habits, 3 subgroups remained: 120 with long near work time, 347 with less outdoor activities, and 253 with little sleep.DNLE was applied to ocular parameters (central corneal thickness, axial length, corneal curvature). Global networks were constructed using Pearson correlations ≥0.6. Local networks centered on each parameter calculated entropy scores integrating standard deviation and Pearson correlation coefficients (PCC). Critical points were identified via stage-specific DNLE scores. Dominant Dynamic Network Markers (DNMs) modules (frequency ≥4 in DNLE) were selected for functional analysis through receiver operating characteristic (ROC) curve evaluation to assess their predictive capacity for accelerated decline in hyperopia reserve.Key metrics included annual hyperopia reserve changes, DNLE scores, and areas under curves(AUC).Critical points emerged 1 to 4 years before myopia onset, with age-stratified patterns. Older children (grades 5-6) showed dual critical points (3-4 and 0-1 years pre-onset), while younger groups (grades 2-3) had single or coinciding critical points. Grade 2 emerged as a critical period where critical points were universally observed across all risk groups. Peripheral refraction, CCT, and AD can serve as DNMs. High-frequency DNMs achieved AUCs ≥0.75 in most groups. Sleep-deprived children exhibited lower prediction accuracy (AUC 0.66-0.69).DNLE successfully detected early warnings of hyperopia reserve depletion, stratified by age and risk factors. Interventions targeting critical windows (grades 2-3) could prevent rapid myopia progression. The method highlights the importance of ocular parameter networks in early detection, offering a framework for personalized prevention strategies.Copyright © 2025 Elsevier Inc. All rights reserved.