Abstract: The determination of dual control indices for concrete pouring temperature based on methods such as extreme probability or maximum entropy lacks a sufficient degree of retention for extreme values in sample data. Moreover, existing literature lacks a clear physical significance for the selection of typical age periods for temperature dual control indices. To prevent the concrete pouring temperature from exceeding the maximum allowable temperature, combined with actual measured temperatures during concrete pouring in a certain concrete dam during hot seasons, this study utilized the Peaks Over Threshold (POT) model to establish dual control indices for concrete pouring temperatures during the semi-mature age period under adiabatic temperature rise conditions. To improve the reliability of the results, a threshold determination method based on the improved 3σ criterion was adopted, and an improved conditional probability determination method for the POT model was proposed. Engineering case studies show that the dual control indices for temperature determined by the improved POT model are close to those determined by the typical extreme probability method. However, since the POT model focuses on the temperature of the pouring chamber under extreme conditions, it provides a more accurate estimate of the probability of extreme events, resulting in more reasonable temperature dual control indices. Additionally, selecting the semi-mature age period of the concrete pouring chamber as the typical age period for determining temperature dual control indices is more conducive to engineering application.