Abstract: Using the winter wheat cultivar Tainong 18 as the experimental material, we analyzed yield stability from 2012 to 2016 under three different treatments: T₁(following typical local field management practices), T₂(high-yield: high nitrogen and water were supplied to foster high grain yield), and T₃(high-yield, high-efficiency: optimized field management including increasing plant density, reducing nitrogen input and delaying of the sowing date). Yield related phenotypic traits, including the number of ears on the main stem and tillers, leaf area index (LAI), photosynthetically active radiation (PAR) interception, dry matter accumulation and distribution, as well as grain yield, were analyzed over four seasons to determine their relationships with annual radiation, accumulated temperature and precipitation. We then determined grain yield stability for each of the three treatments. The amount and distribution of radiation, accumulated temperature, and precipitation varied greatly within each season. The ears on the main stem represented 38.9%, 58.7%, and 66.9% of the total ears, respectively, for wheat grown in the T₁, T₂ and T₃ treatments, indicating that T₁ ears originated mainly from the tillers, T₂ ears from both the main stem and the tillers, and T₃ ears from the main stem. The T₂ and T₁ treatments produced the highest and lowest amount of dry matter and grain yield, respectively. Although having relatively lower dry matter accumulation at maturity compared with T₂, T₃ led to higher grain yield due to high LAI, high PAR interception and utilization, high net canopy photosynthetic rate from booting (especially from 14 days after anthesis) to maturity and a higher harvest index. Among the three treatments, T₃ resulted in the lowest annual range, standard deviation, and coefficient of variation for LAI, PAR interception, and dry matter accumulation. Thus, grain yield was most stable in wheat grown in the T₃ treatment mainly due to stability in biological production during all four seasons.