Thermistor is a very widely used sensor especially for temperature measurement because of its fast response to small temperature changes. The high sensitivity of the thermistor leads to a non-linear behaviour which can give rise to various difficulties such as on-chip integration, direct digital display, wireless capability and so on. So, there arises a requirement for an efficient linearizer to overcome this difficulty. The thermistor is connected to an op-amp signal conditioning circuit (OSCC) which has a stable temperature-voltage relationship over the temperature range 0 °C-100 °C, but suffers with considerable non-linearity of ±12%. In this paper, an adaptive neuro-fuzzy interference system (ANFIS) is used to reduce the non-linearity of the thermistor OSCC. The linearity error is reduced to below ±2% using the proposed methodology and thus making the system suitable to be utilized efficiently for practical applications. © 2006-2019 Asian Research Publishing Network (ARPN).