Stride length estimation (SLE) is a fundamental component of pedestrian dead reckoning (PDR) in indoor navigation and positioning (INP) applications. The knowledge of stride length is crucial for determining the distances covered by pedestrians and estimating their position in real-time. In this study, we proposed a real-time SLE method using innovative textile-based capacitive strain sensors attached to knee pads. The SLE performance of the capacitive sensors was compared with smartphone IMUs, and the results were reported. We applied a supervised SLE approach by creating labeled gait data from participants who wore sensors and walked along controlled paths created with predetermined stride lengths. An adaptive stride detection algorithm was developed to handle data diversity resulting from varying participant characteristics. Furthermore, we investigated the contribution of gait phase features to SLE. The proposed model achieved impressive outcomes with a mean absolute error of 8.73 cm, showcasing its significance in accurate real-time SLE.