Chronic respiratory diseases are conditions that can severely affect lung function as patients grow. The gradual decline in lung function may increase the risk of developing chronic obstructive pulmonary disease (COPD) in adulthood. Therefore, it is important to incorporate pulmonary rehabilitation programs from an early age. The use of games has become an excellent tool to promote long-term treatment adherence. In this context, an innovative joystick design is presented for pediatric respiratory rehabilitation through the development of interactive games. The joystick integrates a differential pressure sensor to monitor the user’s breathing patterns. The centerpiece of this system is the Arduino Nano 33 BLE Sense microcontroller, equipped with a preloaded machine learning model that accurately identifies four key respiratory patterns essential for therapeutic management. Through the control of an avatar, users can complete various challenges by executing these breathing patterns, making the joystick a novel and effective tool for respiratory rehabilitation. The experimental design for real-time identification and recognition of breathing patterns demonstrated a sensitivity of 96.25%, indicating the ability of the system to accurately detect the required breathing patterns, ensuring reliable feedback and improving the effectiveness of the rehabilitation process.