Intelligent mobile terminals (or users) of next generation wireless networks are expected to initiate voice over IP (VoIP) calls using session set-up protocols like H.323 or SIP (session initiation protocol). To guarantee service quality of such applications, the call set-up protocol should be robust against network impairment generated in wireless networks. In this paper, we analyze the performance of the H.323 call setup procedure over the wireless link. We used two call modes of operation for this analysis namely regular and fast connect. The proposed model assumes that a transparent radio link protocol (RLP), but without RLP retransmission, be used for VoIP packets while non-transparent (regular) RLP be used for H.323 control packets. Our analytical models prove that the VoIP call set-up performance can degrade significantly even for moderately high air-link frame error rates (FER). Furthermore, the call setup performance can be improved significantly using the robust radio link layer such as RLP with the modifications proposed in our paper. The analytical results were validated by our experimental measurements.