Abstract: The Yellow River Estuary National Park Candidate Area is a highly representative coastal estuarine wetland ecosystem in China, located in the core area of the newly formed wetlands of the Yellow River Delta. Tidal creeks on tidal flats serve as important corridors for the exchange of materials, energy, and information between land and sea, performing functions such as drainage and sediment transport, nutrient cycling, and support for biological migration. However, most existing remote sensing studies have relied on 30-meter resolution Landsat imagery or single-period observation data. Notable limitations remain in the identification of terminal fine creeks and the assessment of spatial stability at the patch scale. In this study, 28 Sentinel-2 images at 10-meter resolution spanning 2016—2025 were selected, covering low, medium, and high tide levels, with rigorous cloud masking applied. A DeepLabV3+ model was trained on 413 visually interpreted sample patches to extract tidal creeks. Multi-temporal frequency overlay analysis was employed to classify the spatial stability types of tidal creeks. A Comprehensive Landscape Value Index (CLVI) was constructed based on seven equally weighted normalized indicators across three dimensions: scale, morphological typicality, and branching development. Furthermore, a two-tier conservation framework consisting of core conservation zones and adaptive buffer zones was proposed. The results are as follows. (1) The annual extracted tidal creek area ranged from 20.82 to 42.81 km², with a 10-year average of 30.24 km², showing significant interannual fluctuations. The number of tidal creek patches fluctuated between 49 and 91, with no significant monotonic trend. The patch count was 49 in 2016 and 63 in 2025. (2) The α, β, and γ indices varied within narrow ranges of 0.010—0.024, 1.016—1.045, and 0.339—0.348, respectively, indicating that the tidal creek network as a whole approximates a dendritic structure. (3) Pixels identified as tidal creeks in at least 15 of the 28 images were delineated as stable core zones, covering an area of 12.6 km², accounting for 14.2% of the total extracted extent. More than 80% of the extracted extent belonged to intermittently identified areas. Based on 10-year temporal overlay classification, the high-persistence zone covered 16.1 km² (18.2%), the moderate-fluctuation zone covered 17.3 km² (19.5%), and the high-sensitivity zone covered 55.1 km² (62.2%). The spatial overlap index between the two sets of stability zone delineation results was 0.69. (4) The CLVI in 2025 ranged from 0.000 to 0.850, with a mean of 0.236. Only two patches scored above 0.5, and the overall distribution exhibited a significant right skew. Both high-value patches were located in intermittently detected areas, suggesting that a higher degree of tidal creek morphological development does not necessarily correspond to greater spatial stability. Based on these findings, this study delineated 34 km² of Class Ⅰ core conservation zones and 64.4 km² of Class Ⅱ adaptive buffer zones, and established four types of annual monitoring categories: priority conservation, priority monitoring, structural safeguarding, and general management. The "extraction-analysis-assessment-zoning" workflow constructed in this study can provide a quantitative basis for the conservation and adaptive management of tidal creek landscapes in the Yellow River Estuary National Park Candidate Area.