Chiller sequencing control plays an important role in achieving air-conditioning system energy efficiency since chillers usually account for major part of system overall energy consumption in commercial buildings. This paper compares and evaluates the in situ performances of three commonly used chiller sequencing control strategies. These three control strategies are operating online in parallel on a management and communication platform in a super high-rise building in Hong Kong. The comparison and evaluation focus on the perspectives of chiller switch number, electrical peak demand and overall energy consumption in one week. Meanwhile, the different impacts of two systematic errors (i.e. positive errors which cause direct measurement larger than true value and negative errors which cause measurement smaller than true value) on each control strategy are also analyzed. The performance comparison demonstrates that the strategy, using fused cooling load measurement, achieved the best performance. The applications of the online computed chiller maximum cooling capacity and the advanced data fusion technique, which removes the negative impacts of sensor measurement uncertainties, can significantly improve the performance of chiller sequencing control. (c) 2013 Elsevier B.V. All rights reserved.