Recognition of human body posture from the signals of wearable sensor has attracted a great interests in many applications, such as health care, aged care and sports. Main aim of the presented work is to recognize the basic activities of daily living (ADL) and detection of fall using two triaxial accelerometer located at the chest and thigh of the subject. In the present work, posture and Fall detection especially for elderly is performed using a multiscale entropy (MSE) analysis coupled with fuzzy logic (FL) algorithm. Experiments were performed with two wireless triaxial accelerometer. Real time, acceleration were acquired at ADL and were used for the diagnosis. MSE were calculated from the obtained signals and FL was coupled to model and prediction of posture. The result reflect that the proposed noble method is very useful and effective in detecting the different postures and fall. The proposed technique has been optimized to implement in real-time system.