Vibration detection and measurement of a rotary shaft is essential for the diagnosis and prognosis of industrial rotating machinery. However, the imbalance of a shaft, as quantified through vibration displacement, is the most common cause of machine vibration. The objective of this study is to develop a novel technique through electrostatic sensing for the online, continuous and non-contact detection of the imbalance fault of a rotary shaft. A mathematical model is established to extract useful information about the shaft displacement vibration from the simulated signal in the frequency domain. Experimental tests were conducted on a purpose-built test rig to detect the displacement vibration of the shaft at different rotational speeds (400, 800 and 1200 rpm). A normal shaft and an eccentric shaft were tested with the output signals from the electrostatic sensor analyzed. The effectiveness of the proposed method is verified through experimental tests and computational simulation.