This paper presents a high precision temperature measurement instrument based on quartz tuning-fork temperature sensor (QTTS) using Artificial Neural Networks (ANN). The advantage of QTTS is a great sensitivity which makes possible to determine the temperature with the accuracy of 0.01 °C , but the QTTS based temperature measurement instrument often appears as erroneous temperature reading when using standard polynomial calibration techniques over a large temperature range. For high precision temperature measurement, a new method is presented to compensate non-linearity of QTTS based instrument using non-linearity compensation model using ANN by Levenberg-Marquardt algorithm to settle its non-linear problem. The hardware and software parts of the system are integrated in a PC-based instrument used for operation and calibration. ANN based modelling and correction technique has been evaluated experimentally, followed by experimental results obtained by applying the method to QTTS calibration.