The diversity of data center cooling and the complexity of airflow make it difficult to develop general theory for cooling energy efficiency. Based on entransy analysis, this paper proposes a new approach to improve the energy performance of data center cooling processes by reducing the loss of heat transport ability. Minimum entransy dissipation is proposed as a general principle to optimize heat transport performance and indoor environment of data centers. The influence of random mixing of cold and hot air streams on cooling performance is quantified using entransy dissipation theory. A case study is performed to prove the applicability of entransy analysis on evaluation of data center environment.