This paper develops a real-time capacitance ratio estimation system for a novel MEMS tactile sensor that can provide elasticity measurement of a variety of target objects. The tactile sensor is developed by this team for in-vivo tissue elasticity measurements and can be installed on a hand-held probe or on minimally invasive surgical instruments. The sensor readout consists of two channels of capacitance values and the ratio of these capacitance values needs to be reliably estimated in real-time. The proposed estimation algorithm utilizes a recursive least squares method with adaptive forgetting factors, which provides reliable elasticity measurement of polymer specimens and quickly detects changes in elasticity. Extensive experimental results are presented.