Chiller sequencing control is essentially to determine proper thresholds for switching on and off chillers so as to guarantee that the operating chillers can provide sufficient cooling capacity while not waste energy for a given load condition. Total cooling load-based chiller sequencing control determines the thresholds according to building instantaneous cooling load and chiller maximum cooling capacity, which is in principle the best approach for chiller sequence control. However, one challenge for practical applications is that the measure of the cooling load and the estimate of the chiller maximum cooling capacity are associated with uncertainties. To deal with the uncertainties, a stochastic chiller sequencing control is proposed in this paper, which shows that the uncertainties associated with the cooling load measurement can be well described using Normal distribution and the uncertainties associated with the chiller maximum capacity estimation can be described using Uniform distribution. The switch-on/off thresholds are therefore determined in the framework of statistics. An algorithm to realize the stochastic control is developed. Case studies compare the stochastic control with the conventional deterministic method, and the results show that the proposed method can improve the robustness and flexibility of chiller sequencing operation. (C) 2014 Elsevier B.V. All rights reserved.