MPDU aggregation is not applied to real-time voice traffic (e.g., VoIP) in currently available IEEE 802.11n WLAN implementations considering the strict upper bounds of end-to-end delay and jitter. As a result, when real-time voice and non-real-time heavy load traffics are intermixed the resource utilization and overall throughput become poor due to the overhead produced by individual voice MPDU transmissions. In this paper, we propose QoS-aware Adaptive MPDU aggregation scheduler which applies aggregation to voice traffic and adaptively adjusts the aggregation size based on time-varying end-to-end delay and WLAN contention, and QoS requirements (e.g. less than 150 ms end-to-end delay). Experimental results of the proposed scheme show that the overall throughput increased by 57% when 10 stations generate 64 Kbps PCM voice traffic on 270 Mbps PHY rate for 2 ms transit delay representing campus network communication. For 120 ms transit delay configuration that represents intercity communication, the throughput enhancement was 45 % compared to the existing scheme1.