Monitoring periodic limb movements in sleep (PLMS) is important since it is correlated with people's quality of sleep and several other sleep disorders. The clinically approved method of examining PLMS is polysomnography (PSG) where the sleep of patients are examined in a laboratory with various sensors attached to their body. However, PSG is timeconsuming and expensive for patients and the need for costeffective and comfortable PLMS detection method has not been fulfilled. Accordingly, we propose a PLMS detection framework which utilizes a wearable motion-sensor-embedded band. In this work, we study the location to comfortably wear the device and accurately collect data on a foot. Further, to increase the accuracy of classifying PLMS, we propose the Motion Synchronized Windowing technique which segments the intervals where movements occur. Finally, we classify PLMS by using various machine learning algorithms typically used in the human activity recognition. Our proposed system achieves the accuracy of up to 96.92% in detecting PLMS. Therefore, our system is a costeffective and convenient method of monitoring PLMS.