One of the most important issues in the wearable healthcare sensors is to minimize the motion artifacts in the vital signals for continuous monitoring. This paper presents a reflected type photoplethysmograph (PPG) sensor for monitoring heart rates at the artery of the wrist. Active noise cancellation algorithm was applied to compensate the distorted signals by motions with Least Mean Square (LMS) adaptive filter algorithms, using acceleration signals from a MEMS accelerometer. Experiments with a watch type PPG sensor were performed to validate the proposed algorithm during typical daily motions such as walking and running. The developed sensor is suitable for ubiquitous healthcare system and monitoring vital arterial signals during surgery.