Optical tweezers is the probe studied weak force measurement， Measurement Range and Measurement Accuracy of force are important index of the system of optical tweezers. BP optimizations for neural networks to correct the non—linearity of calibration of the force of optical trap are provided in this paper． The range of measurement of the system of optical tweezers of data is improved without increasing the complexity of hardware. With the comparison of BP method and curve fitting method

it is presented that the nonlinear correction method based on neural network has high measurement precision. Experimental results of Nano-Optical-Tweezers system show that in such a way

the performance of the system is effectively improved.