In this paper, we propose a facial direction estimating system from low resolution facial images captured by surveillance camera. The proposed system first detects pedestrians by background subtraction, then trimming head area by template matching and finally estimating facial direction by using support vector machine (SVM). Experimental results show that, when we use 8times9 pixels facial images as input, training image resolution and estimation image resolution are 8times9 and estimation directions are in 13 directions by the face databases, the estimation rates are more than 90% by our method.