The invention discloses a wearable data collection system based on multi-sensor fusion and a method adopted by the system. The wearable data collection system comprises a mechanical skeleton, wherein the mechanical skeleton comprises a waist bracket (1), thigh connecting rods (2), shank connecting rods (3), ankle joint connecting rods (4) and intelligent shoes (5) which are arranged from top to bottom, wherein a main control box (6) is arranged on the back of the waist bracket (1), and posture instruments (7) are arranged on the main control box (6), the thigh connecting rods (2) and the shank connecting rods (3); the thigh connecting rods (2) and the shank connecting rods (3) are provided with node plates (8); the thigh connecting rods (2), the shank connecting rods (3) and the ankle joint connecting rods (4) are respectively provided with encoders (9). The wearable data collection system and the method are used for accurately measuring the angles of the lower extremity joints and the distribution condition of the pressure of the foot soles when a wearer wearing an exoskeleton robot walks, so that whether the exoskeleton is positioned in the stable walking state or not can be judged, and the current health condition of the wearer can be evaluated.