||The advancement of technology coupled with the growing demand for personal communications has fueled a recent increase in interest in the study of wireless communications, which is proving to be the most remarkable development in telecommunications since fiber optics. There is widespread belief that consumer demand for wireless communications will continue to expand for the foreseeable future. However, many obstacles still exist in the implementation of wireless personal communications such as interference, security and multipath issues. With the increased need for communications and with a "fixed" available spectrum, it has become necessary to use the spectrum more effectively. Spread spectrum is a technique for efficiently using the spectrum by allowing additional users to use the same band as other simultaneous users. Code division multiple access (CDMA), which is based on spread spectrum, has been proposed as the next generation multiple access technique for wireless personal communications. The choice of CDMA as a multiple access technique is attractive for many reasons. Its higher capacity potential and its privacy and multipath rejection capabilities are just a few. However, due to the inherent complexity of CDMA systems, the performance analysis of such systems is often intractable and computationally difficult. The performance capabilities of CDMA systems are determined by a variety of methods including analytical studies and computer simulation. Theoretical analyses may require idealistic assumptions which make these approaches either inaccurate or intractable when practical systems are considered. On the other hand, simulation techniques can incorporate all system complexities much more effectively than analytically-based methods while at the same time maintaining their precision rather than sacrificing it through simplification. This thesis deals with the performance analysis of direct sequence (DS) CDMA systems. In particular, it studies simulation-based techniques namely Monte Carlo and Importance Sampling simulations. Although Monte Carlo simulation is a useful method for analyzing the performance of communication systems, it is inefficient for estimating the low error rate performance which is usually the issue of interest. Importance Sampling is a modified Monte Carlo method which can reduce the computational cost of simulations by several degrees. However, to achieve a desired accuracy and make the Importance Sampling approach more efficient proper care must be taken to design the scheme which involves biasing the original system. In this work, it is shown that the selection of biasing strategies is of paramount importance to the performance of Importance Sampling. Furthermore, the results of Importance Sampling will be compared with ordinary Monte Carlo and theoretical results to prove the potential and effectiveness of the Importance Sampling technique when applied to the simulation of CDMA systems.