We propose an efficient scheme for the call admission control in CDMA communication system. A large number of papers in recent literature have been dedicated to the CDMA call admission control. In most of these papers, Poisson traffic model is assumed. However, the recent measurements of local-area and wide-area traffic have shown that network traffic exhibits variability at a wide range of scales-self-similarity. It has been reported that the long-range dependence aggravates the system performance if it is taken into consideration and the traditional Poisson model cannot capture the feature of non-summable autocorrelations. In this paper, we examine a mechanism that gives rise to self-similar network traffic and present some of its performance implications. The mechanism we study is the transfer of files of messages whose size is drawn from a heavy-tailed distribution. We study call admission control algorithms for improving performance under self-similar traffic conditions. Although the scale-invariant burstiness imposes a limit on the ability to achieve high quality of service (QoS) and utilization jointly, the long-range dependence associated with self-similar traffic leaves the possibility that the correlation structure may be exploited for performance enhancement purposes. We show that a significant improvement in performance can be obtained if proposed CAC is active and that the relative performance improvement increases as long-range dependence is increased.