In image analysis, horizontal integral projecting and vertical integral projecting algorithms are usually utilized to describe the gray distribution of targets. If targets rotate, it is difficult to find the relationship among the projecting curves of the same targets in different orientations. In order to overcome this default, two kinds of statistic methods called circular integral projecting and circular grads integral projecting are proposed based on circular projecting. Furthermore those methods are compared with horizontal integral projecting and vertical integral projecting. The experimental result shows that the curves corresponding to circular integral projecting and circular grads integral projecting own excellent steadiness when target rotates and flexes.