Sarcopenia disease is due to low muscle mass in humans. Sarcopenia leads to osteoporosis, metabolic syndrome and difficulty in performing day-to-day activities. At present, Dual-energy X-ray Absorptiometry (DXA) measures muscle mass with few limitations. They are variations in measurements according to region under investigation, irregularities in hydration status, and low precision in tall and obese persons. These limitations are due to low dosage level of X-ray radiations in certain muscle regions of human body such as heart, head, lower and upper extremities. This paper presents a non-invasive passive flexible Ultra Wide Band (UWB) Myogram antenna sensor for the prediction of Sarcopenia through human muscle mass measurement. This antenna is adhesively fixed on ventral surface of forearm and biceps for the measurement of skeletal and lean mass respectively. The proposed antenna sensor performs electromagnetic energy absorption from muscle tissues under radiating near-field condition. The muscle tissue signal from antenna is applied to blind source filtering-Non-negative Matrix Factorization (NMF), then subjected to Multi-Synchro Squeezing Transform (MSST), and finally correlated using linear regression machine learning algorithm to diagnose Sarcopenia. Furthermore, the proposed methodology is developed as a product through the MATLAB Mobile App compatible with Android devices. The proposed method of diagnosing Sarcopenia achieves an accuracy of 85% in fifty samples.