Real-time temperature monitoring data quality is important for intelligent concrete cooling control. This paper presents a moving-average method for real-time concrete temperature monitoring data which extends the traditional moving average to multiple dimensions including time, space, and quality. Dynamic analyses of the raw data average in the intelligent concrete cooling control system provide temperature data that characterize the conditions at a given moment, spatial location, or object, which significantly reduces the effects of systematic and random errors in the raw data and reduces data distortion caused by instrument failures, sudden environmental changes and other effects. This method then reduces the impact of data errors on subsequent data calculations, simulations, and analyses. Analysis of real-time concrete temperature monitoring data from the Baihetan and Wudongde arch dams shows the advantages of this method for improving the monitoring system, data quality, and decision-making ability of management personnel. The results then provide a reference for concrete temperature control system design and quality evaluations.