Liquid level is an important parameter, which is measured in various industrial processes. However, most existing methods are invasive. This paper presents a design for measurement of liquid level of horizontal pipes, based on electrical capacitance tomography (ECT) with a 4-electrode ECT sensor, which may be conducted by a robot hand. This approach enables generation of cross-sectional images to visualize liquid distribution and measure the liquid level. Even when the liquid surface is inclined at an angle to the horizontal, the corresponding liquid partition can be determined by using the K-nearest neighbors (KNN) algorithm. The main advantage of this liquid-level measurement method is its characteristic of non-invasive and non-intrusive measurement. It provides visualized information, facilitates multi-point measurements in a pipeline system, and can effectively avoid human involvement in high-risk environments.