Real-time identification of frequency-hopping millimeter-wave (mm-wave) signals is a real challenge, due to the high demand for detection bandwidth and processing speed. In this paper, we propose and demonstrate a novel microwave photonic approach to identifying frequency hopping of mm-wave signals based on the concepts of photonic time stretch (PTS) and Reservoir Computing (RC). The PTS scheme allows the modulated signal to be slowed down hence reducing the required detection bandwidth. The developed RC model offers unique features such as being more efficient and time saving in temporal data pattern classification than traditional methods. According to the simulation, RC can recognize the hopping instants and give a precise hop timing estimation in real-time.