The estimation of average lifetime has been considered as a serious reliability concern for VLSI components. However, with the tremendous downscaling of MOSFET devices, it becomes very difficult to predict such measure with satisfactory precision due to the amplified parasitic effects at the deep scale level. The present work aims at investigating the efficiency of adaptive neuro-fuzzy inference system approach as a predictor of junctionless gate all around MOSFET device lifetime. Our predictions are compared versus the numerical results obtained by ATLAS simulator, where a good agreement is obtained.