Autofocusing is an important technology in the measurement of images and computer vision. After careful experiments, it is found that some conventional autofocusing algorithms are valid without the influence of the noise. However, they will be invalid in the influence of noise. So, an auto-adaptive algorithm to autofocusing is proposed. At first, edge pixels are separated from nonedge pixels by an adaptive threshold in small local area. The global threshold is adapted to eliminate the influence of noise further. Finally, images are computed with some autofocusing method carefully, which are calculated by the adaptive method. At the same time, the focusing curves are plotted. Experiments show that the method not only owns the characteristics of good unbias, clear single apex and high sensitivity and so on, but also has strong performance of antinoise.