Third-Generation based Wide Band Code DivisionMultiple Access (W-CDMA) networks have been characterized for high data rate services accessible to a large number of users. But multipath and co-channel interference puts a limit on system capacity, throughput and new multimedia information services like wireless internet or video streaming. Smart antenna technique that includes Direction of Arrival (DOA) approach and Adaptive Beamforming algorithm is a promising means to improve the over-all system capacity and to overcome limited spectrum problem by effectively reducing the multipath and co-channel interference. In this paper, DOA estimation is done by Multiple Signal Classification (MUSIC) algorithm and Minimum Variance Distortionless Response (MVDR) algorithm is used for adaptive beamforming that continuously adjusts the weight of antenna arrays for creating a beam to track the desired user automatically, and to place null towards the interference user. Here, we improve on the MVDR method in order to assign the weightings. Instead of using a single linear equality constraint, a variable constraint has been used to obtain the optimum weights that decrease the gain in the undesired direction and desired direction comes out with more gain and without distortion. Obtained results are analyzed for the conventional method and the proposed method by observing beamforming spectrum by first varying number of antenna elements and then by varying the spacing between elements. For both the cases, the performance is better for the proposed method in terms of lower noise floor, more relative power and sharper peaks in the desired user direction than conventional method.