Disclosed is a method for encoding a video signal through discrete cosine transform (DCT) and motion estimation (ME) and an apparatus therefor. The method for encoding the video signal simplifies the ME with reference to DCT coefficients. In a method for estimating motion in a video frame compression system using DCT, flatness degrees of the blocks is determined according to the number of DCT coefficients having a non-zero value among DCT coefficients transformed in units of blocks. A reference frame is formed by recovering video frame data from some or all of the DCT coefficients corresponding to the flatness degrees of the blocks. Precision of motion estimation (ME) for a current macro block (MB) of a current video frame is dynamically changed corresponding to the flatness degree of the reference frame.