In this paper, an efficient variable transmission frame length (VTFL) speech coding method is proposed and its performance is studied by computer simulation. The proposed speech coding method is based on the idea that the performance of speech coding would be improved by varying the transmission frame length according to the stationarity of input speech signal and vector quantizing the representative feature vector of the transmission frame. In the proposed speech coding method, the feature vector sequence consists of PARCOR coefficient vectors obtained by analyzing input speech signal sample-by-sample using the prewindowed recursive least square lattice algorithm. We take the segmentation from the input speech signal and then determine the representative PARCOR vector in each segment. By joining the consecutive segments of phonetically similar characteristics using the likelihood ratio distortion measure, we can obtain transmission frames of variable length. From the computer simulation results, the proposed VTFL speech coding method yields the reduction of total transmission bit rate while maintaining the good reproduced speech quality.