In the use of auto-focusing technology based on image processing of aerial camera, the precision decreases due to the influence of image′s noise. Currently the relationship between signal to noise ratio (SNR) of image and auto-focus precision is mainly based on the experimental results and subjective judgments, but there is no universal theory and exact conclusion on them. The relationship and an exact mathematical expression are analyzed and given between the SNR and auto-focus precision, experiments show that the SNR of image has a positive correlation to auto-focus precision, the higher SNR of the image, the higher precision of the auto-focus precision. Through the standard of the expectation and variance of focus measure, the influence of SNR on precision is analyzed, and results show that the influence of noise can be weakened by calculating the focus measure more times. Experiments are designed and the changes of SNR with variance of focus measures are reconded by applying the same focus method in different conditions. Results show that when the SNR of image rises from 27 dB to 39 dB, the variance of focus measure function decreases from 0.262 to 0.0611, meaning that auto-focusing accuracy is improved by about 30%. The result proves the correctness of the theoretically analysis.