Accumulated temperature is an important factor for modeling crop growth. It is stable in theory, but in practice, the accumulated temperature needed for crops during different growth stages differs markedly among different years, regions, and varieties, so the stability is relative and the instability is absolute. Therefore, it is useful to establish a general model to calculate accumulated temperature that is relatively stable and applicable to different maize varieties. In this study, we analyzed the stability of accumulated temperature and the parameters of the non-linear accumulated temperature model (NLM). The NLM was optimized to improve its application range. A linear accumulated temperature model (LM) was also optimized based on the most important factor affecting the stability of accumulated temperature. We compared different methods for calculating accumulated temperature. We found that the accumulated temperature needed for crops during different growth stages differed markedly among different years, regions, and varieties. The main reason for the instability of calculated accumulated temperature values was temperature strength for a certain variety. Therefore, the calculation method was revised by adding a quadratic function, generating the temperature revision model after revision (TRM). The parameter Q of the NLM is a thermal-sensitive parameter. There were strong correlations between Q and mean active accumulated temperature or effective accumulated temperature for different varieties during emergence to maturity, indicating that Q was related to the maturity type. Consequently, we proposed two general accumulated temperature models, AARM and EARM, in which the parameters of NLM were denoted by the active accumulated temperature or effective accumulated temperature. Comparing the different models, the TRM generated minimal bias but AARM and EARM had a wider application range for many varieties on a large scale. AARM had better simulation effect, while EARM was more stable. The applicability of the optimized models was improved. The results provide a new approach for optimization of agrometeorological indexes and upscaling of accumulated temperature models for other crops.

积温是模拟作物生长的重要因素。它在理论上是稳定的，但在实践中，不同生育阶段，不同年份，不同地区和不同品种的作物所需的积温明显不同，因此，稳定性是相对的，并且不稳定是绝对的。因此，建立一个通用模型来计算相对稳定并适用于不同玉米品种的累积温度是有用的。在这项研究中，我们分析了累积温度的稳定性以及非线性累积温度模型（NLM）的参数。对NLM进行了优化，以扩大其应用范围。线性累积温度模型（LM）也基于影响累积温度稳定性的最重要因素进行了优化。我们比较了计算累积温度的不同方法。我们发现，不同生长阶段，不同年份，不同地区和不同品种的作物所需的累积温度明显不同。计算出的累积温度值不稳定的主要原因是某个品种的温度强度。因此，通过添加二次函数来修改计算方法，从而生成修改后的温度修改模型（TRM）。参数 通过添加二次函数来修改计算方法，并生成修订后的温度修订模型（TRM）。参数 通过添加二次函数来修改计算方法，并生成修订后的温度修订模型（TRM）。参数NLM的Q是热敏感参数。Q与成熟期不同品种的平均有效积温或有效积温之间存在很强的相关性，表明Q与到期类型有关。因此，我们提出了两个通用的累积温度模型，AARM和EARM，其中NLM的参数由有效累积温度或有效累积温度表示。比较不同的模型，TRM产生的偏差最小，但是AARM和EARM在大规模的许多品种上具有更广泛的应用范围。AARM具有更好的仿真效果，而EARM更稳定。优化模型的适用性得到了改善。研究结果为优化农业气象指标和提高其他作物的积温模型提供了一种新方法。