Force control is very crucial in nanoimprint lithography (NIL). It is necessary to develop a high-performance force sensor to provide real-time force feedback for the control process. Due to the unique procedure of NIL, the developed force sensor should include a high sensitivity, broad bandwidth, and large measurable range. However, these characteristics are normally conflicting in nature and cannot be physically avoided by any force transducers so far. To address this problem, this paper presents a novel dual-range force sensor, and uses a heuristic multiobjective optimization method to make a tradeoff among these characteristics. This method is based on the particle swarm optimization algorithm, meanwhile employs the Pareto ranking scheme to find optimal solutions. Through proper optimization, not only the three characteristics are compromised, the lowest stress concentration of the sensor body is maintained as well. To demonstrate the effectiveness of the optimization, numerical simulations with finite-element software COMSOL are conducted. A prototype sensor is then fabricated according to the optimization results. The simulation and prototype test results indicate that the optimized sensor has a resolution down to 800 mu N, a bandwidth up to 150 Hz, and a measurable range up to 180 N. All the results prove that the developed force sensor possesses a good property for high-performance force measurement, and satisfies the needs of NIL as well.