Today air conditioning system, the control method of the chiller and cooling tower water cooling towers mostly fixed fan speed or with the drive to achieve different operating modes. The most common operation modes is fixed approach temperature or depend on the temperature of cooling tower water to control the cooling tower fan speed. Many manufacturers cannot make the correct parameter settings under different load patterns, often with experience as a basis for setting the parameters. Therefore, parameters on energy consumption have become an important objective of this study. Data mining is a method of use of database analysis performed. Through large amounts of data, and extract the hidden past is not known but credible and effective knowledge. On the other hand, according to the parameters set by the user, to find information of interest to the user from data in a group of unprocessed. The data is provided for decision-makers to reference for judgment as a basis after data transformation and post-processing. This paper intends to use data mining - association rule approach to building electricity database for analysis. With the establishment of minimum support and minimum confidence level, we can get the most suitable energy and the approach temperature of operation for chiller and cooling tower systems in buildings.