We present a simple and efficient scheme to predict the number of bits needed to code a DCT block with variable length code (VLC). Utilizing this new approach, a massive computing effort is averted in rate-constrained video encoding process. The scheme defines an easily computable parameter that has a strong correlation with the number of bits needed to code the DCT data. A second order polynomial is developed to approximate the correlation and is used to predict the number of VLC bits. This approach is further used to streamline the encoding/transcoding process for DV, and has a potential to be used in other DCT/VLC based coding schemes to facilitate rate control.