Various sensors have been recently integrated in a smart-phone and step detection method based on the acceleration sensor of a smart-phone has been introduced for indoor positioning scheme. Many researchers have put their interests on the study of step detection algorithm based on the acceleration sensor. However, the estimation performances of these methods are not good enough or just suitable for the situation when the position of the smart-phone is fixed in hand because the performance is significantly degraded in other cases. In this paper, a novel step counting algorithm based on the acceleration and the gravity sensors is proposed to enhance the estimation performance regardless of the position of a smart-phone and the motions of a pedestrian likewise walking or running. The effectiveness of the proposed scheme is demonstrated with experiments and the performance of the proposed scheme is compared with those of the conventional schemes. According to the results, the performance of the proposed scheme is enhanced compared to the conventional schemes no matter what the pedestrian is walking or running although the phone is in different places like trouser pocket, shirt pocket and hands.