Abstract: The stress state at high arch dam heel is an important index for evaluating dam safety, but its variation and influencing factors are complicated due to its statically indeterminate structure, concrete placement during dam construction, changes in water level during reservoir operation, and changes in the internal and external temperature of the dam. To determine the dam heel stress, in-situ monitoring and numerical analysis are two main methods, but they often produce significantly different results. This paper discusses the difference between these two methods in dam heel stress evaluation. For in-situ monitoring, through the analysis of the working principle of the compressive stress meter and strain gauge, we suggest the evaluation of effective stress at dam heel, and an improvement on the arrangement of osmometers, compressive stress meters and strain gauges. For numerical analysis, we examine the relationship of dam heel equivalent stress versus mesh size along dam thickness, and give suggestions for effective mesh partition in calculation of dam heel stress. Results show that although both methods have their own influencing factors, the improved monitoring method combined with the numerical calculation of high accuracy can give a satisfactory evaluation of dam heel stress.