Wearable robots are devices that users wear to assist with muscle strength or overcome physical limitations. Research on wearable robots is being conducted extensively both domestically and internationally. Gait motion is influenced by various factors and varies significantly from person to person, making analysis and prediction complex. Furthermore, the sensors and algorithms used in wearable robots must be minimized to achieve lightweight and simplified conditions. To solve these problems, we propose a GCP algorithm control model for walking assistance. This control algorithm measures the Gait Cycle Percentage (GCP) in real time based on IMU sensor data attached to both thighs and then provides the appropriate assistance force. For stability and safety, no control is performed if the gait is too fast or too slow. To ensure stable data, we employ a buffering concept, which filters out the most unusual data from the previous three data points and takes the average of the remaining two. This lightweight algorithm can meet the mainboard's requirements for low weight and cost. Furthermore, with memory savings, it enables the addition of high-level algorithms such as Human-in-the-loop (HITL), which provides a future platform for personalized assistive profiles.