In this thesis, we propose a pulse oximetry-based $SpO_2$ and heart rate detection algorithm for mobile devices. Because strong motion artifacts occur during exercise, it is needed to develop a new artifact-robust algorithm for $SpO_2$ and heart rates. However, the researches related to pulse oximetry are mostly on the use of hospital care systems. In this research, we analyze the effect of motion artifacts to the pulse oximetry algorithm and develop a new artifact-robust pulse oximetry algorithm for people in exercising. This research deals with two subalgorithms. One is related to $SpO_2$ detection, and the other is the algorithm for HR detection. In the case of $SpO_2$ detection, we reduced processing time of Masimo’s Discrete Saturation Transform(DST), which is the most widely utilized to $SpO_2$ detection in artifact environments, for mobile devices. We achieved 91.31% of error reduction with the proposed $SpO_2$ detection algorithm. In the case of heart rate detection, the performance improvement is realized by the average magnitude difference function and median filtering for artifact robustness. The proposed heart rate detection method achieved 88.23% of improvement in an error reduction sense compared with the conventional method based on bandpass filtering. Experimental results confirm that the proposed algorithms can be sufficiently applied to mobile healthcare systems during exercise.