Nano-displacement measurement is one of the most important aspects of nanomanipulation. However, the narrow working space and the heat sensitivity of the microscope limit the installation of many displacement sensors. The self-sensing and time-digit-conversion (TDC) method can overcome the above limitations, making these two methods ideal for nano-displacement measurement. The former method has a high sampling frequency, but its accuracy is low. In contrast, the TDC method is more accurate, but its sampling rate is low. To solve these problems, a fusion measurement method was proposed, thus allowing us to combine the results of the self-sensing and TDC. Specifically, an improved Kalman filter was used to overcome the asynchronous multi-rate problem. Moreover, we fully utilized the information of the calibration instrument using the neural network. As for the overfitting problem, we adopted a neural network with convolution filtering. Our method achieved a precision of 47.9% higher than the traditional method, as well as a linearity (R2) of 0.99990 throughout 3,500 nm range.