The present invention relates to super-frame sound track parameter quantizing method and belongs to the field of low rate phonetic compression and coding technology. The method includes first dividing input phonetic signal sampling points in time order into frames and constituting one super-frame with N continuous subframes; extracting sound track A parameter from each current super-frame, converting into LSF parameter and selecting M subframes from each super-frame as the direct quantizing frame; quantizing the LSF parameters of the M subframes in multistage matrix quantizing process; interpolating with the LSF parameters of the direct quantizing frame to obtain the LSF parameter of un-quantized subframes in current super-frame; and cnverting the quantized line spectrum parameter into sound track A. The method can raise the quantizing precision of sound track parameter and suitable for phonetic coding of 300-600 bps or lower rate.