Non-invasive sensors are designed to offer ease of installation and avoid contact with dangerous hot wires. To optimize the array design of sensors, a multi-objective optimization method considering sensitivity, array arrangement, and cost is proposed in this study. The total array induced charge Q and the location sensitivity S1 are established to evaluate the sensor performance. On the basis of these evaluation indexes, design criteria for the array arrangement are proposed, which include the electrode shape, the array profile, and the space interval angle between electrode elements. Following this design basis, two objective functions for the array design are investigated. The first objective function is based on the entropy-CRITIC weight of the two regression models and the second objective function is based on the manufacturing cost. For the multi-objective design, the NAGA-Ⅱ algorithm is employed and the VIKOR strategy is used to select the decision from the Pareto set. Optimization results show that, compared with the original design, the value of the first objective function is improved by 74.35% and the value of the second objective function is reduced by 22.99%. It is found that the rectangular shape is suitable for the single-electrode element design. The sensing array is best designed in a ring profile, and the optimal number of electrodes is 3.