The present invention discloses a low-complexity method of intraprediction mode selection, two groups of sample points representing texture characteristics are extracted from the current block, according to the cost values of adjacent blocks, a threshold is calculated. The method includes four steps; firstly, the method judges if the current block is located at the special position of an image frame, if yes, an optimal prediction mode at the special position is calculated and then exported, and calculation is finished; secondly, if not, the cost value of the current block under the most possible prediction mode is calculated, if the cost value is less than the threshold, then the mode is exported, and the calculation is finished; thirdly, otherwise, a mode with the least cost value is selected from all non most possible prediction modes and used as a temporary optimal prediction mode; fourthly, the cost value of the most possible prediction mode is compared with the cost value of the temporary optimal prediction mode, and the mode with the least cost value is used as an optimal prediction mode and exported. Because the present invention selects a characteristic point to replace the whole data block, on the premise that the video encoding quality is not obviously affected, the present invention notably reduces the calculation load of the intraprediction mode selection and increases the real-time performance of video encoding.