In this thesis we developed a resource allocation framework for wireless code division multiple access (CDMA) networks that support multi-class traffic with different data rates and bit error rate requirements. We proposed a new resource allocation scheme using joint adaptive power control and antenna array multiuser receiver in multipath fading system. In this scheme both transmit power and receiver filter adapt to time-varying fading channel state. By dynamically assigning users appropriate transmit power and receiver filter, the scheme can guarantee users' diverse quality of service (QoS) requirements and significantly improve quality and capacity of the system. We derived theory for abstraction of bandwidth resource and characterization of system capacity for multi-class traffic in multipath fading system. Bandwidth resource allocated to a user can be abstracted as "effective bandwidth", determined by the user's source data rate and target bit error rate. Capacity in multipath fading system can be characterized by a theoretical bound. Simulations show that actual system capacity in multipath fading environment is close to the theoretical bound at large power constraint. The multirate multiuser receiver can achieve significantly larger capacity for users with higher source data rate and lower target bit error rate than conventional matched-filter receiver. The antenna array multiuser receiver can provide large capacity for users with higher signal-to-interference ratio requirement and achieve high bandwidth utility at lower power constraints. We also proposed a new call admission control scheme for CDMA cellular system that incorporates adaptive power control and antenna array multiuser receiver and supports multi-class traffic. The scheme is based on estimation of other cell interference and implemented distributively in each cell. The scheme can adapt to various traffic distribution. Capacity determined by the admission control agrees well with the actual system capacity in the simulation. The scheme can achieve high bandwidth utilization and guarantee QoS requirements of all the users.