[Objective] To address the mismatch between traditional rainfall analysis methods (April to October) for hilly irrigation areas in south China and actual intra-seasonal water demand during rice growth stages (late rice from July to October), this study focuses on intra-seasonal rainfall during the rice growth stages, integrating GIS technology to investigate the water supply reliability of the Maweizao irrigation area. [Methods] Using daily meteorological data from 1989 to 2019 in the irrigation area, the intra-seasonal rainfall frequency analysis method was employed to identify typical representative years and characteristic values for normal years (P=50%), moderately dry years (P=75%), and dry years (P=90%). The FAO Penman-Monteith method and water balance method were then used to calculate the crop water requirements and net irrigation water requirements for rice. [Results]The results showed that: (1) In dry years, the intra-seasonal rainfall during the late rice growth stages (160 mm) accounted for only 21.2% of the total rainfall from April to October (755 mm). Moreover, a mismatch was observed between the rainfall peak (August) and the critical water demand period (booting to heading stage, September). This led to 14% higher net field irrigation water requirements (560 mm) calculated by intra-seasonal rainfall frequency analysis compared to traditional methods, accurately reflecting the typical contradiction in hilly irrigation areas where there was “no rain during water demand periods but excessive rain during non-demand periods.” (2) GIS-based spatial simulations revealed a distinct bimodal structure in the irrigation area during dry years. Croplands near the main water source (Maweizao Reservoir) benefited from sufficient storage capacity (27.02 million m³) and a canal system integrity rate above 85%, achieving a water supply reliability rate greater than 80%, thus forming a high-yield and stable-production core zone. Areas dependent on small reservoirs for water regulation and storage, where storage capacity utilization declined to 60% due to sedimentation, had a water supply reliability rate of 60%-80%. Limited by scattered ponds (406 ponds), insufficient catchment areas (<5 km² per pond), and damaged main and lateral canals (integrity rate <40%), the overall reliability rates dropped below 40%, posing a high risk of yield reduction. (3) For every 10% increase in water supply reliability rate, late rice yield increased by 35-50 kg per mu(1mu≈666.67 m²), showing a significant positive linear correlation (R²=0.89). When the reliability rate exceeded 80%, soil water content remained stable at 18%-24% (optimal range for rice growth), resulting in yields of 400-500 kg per mu.When the reliability rate fell below 40%, soil water content dropped sharply below 10%, leading to plant wilting or even total crop failure (yield <200 kg per mu). Within the 60%-80% range of reliability rate, each 1 m³ irrigation water increase produced an extra 1.2-1.5 kg of rice, indicating optimal resource use efficiency. [Conclusion] By focusing on intra-seasonal rainfall during rice growth stages, this study reveals the underlying mechanism of irrigation water supply-demand imbalance in hilly irrigation areas and proposes the following three practical strategies. Over 70% irrigation water should be allocated during the booting to heading stages (September) based on crop water requirements, with priority given to areas maintaining water supply reliability rates above 60%. For areas with water supply reliability rates below 40%, the “pond desilting + intelligent water control” project should be implemented to increase small water source utilization rate from 45% to 75%, while restoring main and lateral canals to achieve an integrity rate above 60%. By focusing on intra-seasonal rainfall during rice growth stages, this study provides a scientific basis for precise irrigation management and confirmation of agricultural water use rights in hilly irrigation areas, holding important practical significance for optimizing water resource allocation and enhancing grain production capacity.