Abstract: High-temperature heat damage has been a serious risk to the rice water use efficiency (WUE) under global climate change. Critical thresholds can be required to identify in sustainable agriculture and water resources. For instance, the middle and lower reaches of the Yangtze River—a key rice-producing region where output contributes over one-third of the nation's total rice yield—experiences frequent high-temperature heat damage events that severely threaten the stability of rice production. Therefore, this study aims to calibrate and validate the ORYZA V3 crop model. Soil properties, crop data, field management data, and historical meteorological data were collected from two typical experimental stations in Nanjing and Kunshan within the middle and lower reaches of the Yangtze River. Then, 126 high-temperature heat damage scenarios were set (6 stage combinations × 3 durations × 7 temperature ranges) covering single-stage (booting, flowering, and grain-filling), multi-stage (booting+flowering, booting+grain-filling, and flowering+grain-filling), different durations (D1–D3), and varying temperature ranges (T1–T7), particularly without altering existing cultivation and agronomic measures. According to the calibrated and validated ORYZA V3 crop model, a systematic investigation was made to simulate the impacts of each scenario on the rice reproductive growth at two typical experimental stations. A mechanistic analysis was then implemented to clarify how high-temperature heat damage affected the rice yield, growth period, water requirements, and WUE. Threshold responses were identified to propose the adaptive strategies for the national food security and sustainable agriculture. The results indicate that: 1) Model adaptability: There was an excellent correlation on the evaluation metrics during both the calibration and validation periods. Specifically, the R² generally ranged between 0.89 and 0.99. The relative root mean square error (RMSEn) for the total dry weight (WAGT), weight storage organs (WSO), stem weight (WST), leaf weight (WLVG), and leaf area index (LAI) during calibration (validation) periods were 6.20% (7.50%), 14.91% (8.82%), 14.32% (13.51%), 19.91% (19.92%), and 13.42% (19.03%), respectively. All RMSEn values were below 20%, indicating the high simulation accuracy and strong adaptability of the model to rice growth at the two experimental sites. 2) Developmental Rate Transition Threshold: High-temperature accumulated degree thresholds significantly dominated the developmental progress. The rice growth period shortened in regions with the low thresholds, where HDD ≤ 11.1 °C·d (Nanjing) and 20.7 °C·d (Kunshan). Above these thresholds, both stations exhibited a significant reduction in the developmental rates. For instance, Kunshan shared a 28-day prolongation of the growth period under extreme heat stress. Moreover, the sensitivity to high-temperature heat damage greatly varied over the developmental stages: booting > flowering > grain-filling stage. The booting stage exhibited the highest sensitivity, due to its direct link to the cell cycle regulation of the reproductive meristems. 3) WUE Critical Threshold: Under high-temperature heat stress, the WUE of rice continuously decreased. A nonlinear acceleration of decay was observed when coupled heat stress occurred during the flowering-grain filling stage and HDD ≥40 °C·d. Under extreme heat stress, the WUE declined to 16.99% in Nanjing and 16.80% in Kunshan. 4) Water Requirement and WUE shared Inverse Relationship: Under single-stage heat stress, the increase in water requirement or decrease in WUE was significantly greater in Kunshan than in Nanjing. Multi-stage coupled heat stress further amplified this difference. The water metabolism resilience of rice under high-temperature heat stress in Kunshan was far lower than that in Nanjing. 5) Significant Effect of Multi-stage Coupled Heat Stress: The rice yield exhibited a weak linear decline under single-stage heat stress, where the reduction slope in Nanjing was higher than that in Kunshan. Yet the reductions remained within 15%. Multi-stage coupled heat stress also triggered nonlinear yield collapse. When the HDD ranged from 60 to 112 °C·d, the maximum yield reduction reached 79% in Nanjing and 77% in Kunshan. 6) High-temperature heat stress induced a cliff-like decline in WUE, where the carbon-water coupling balance was disrupted during rice reproductive growth. It was recommended to adopt the irrigation system optimization and planting periods in order to establish an adaptive framework. This finding can provide scientific support to the rice adaptation strategies under dual climate risks (compound drought-heat events) under climate change.