Jitter buffering is a key component in multimedia and real-time services. A jitter buffer is used at the receiver side to control packet rate and to compensate for the delay variation, commonly called jitter, by its function of tradeoff between delay and loss. The packet loss at the jitter buffer can be induced by either underflow and overflow, and we note that buffer underflow probability is an important performance metric for the Quality of user Experience (QoE). Since the estimation of this parameter is difficult for most buffer models, we propose in this paper to show that by controlling the mean delay variation in the buffer, we could control the underflow probability. Thus, we provide an analytical approach for jitter delay estimation in jitter buffer.