A soft sensing strategy for diathermancy of sintered material layer in sintering processes is presented this paper, by a hybrid mode ling of Artificial Neural Networks (ANN) and Fuzzy inference (FI) with compensation of a fine adjusting factor (FAF). Secondary variable selecting, data collecting and processing, modeling and its modifying of this soft sensing have been discussed in order to estimate or identify sintered material layer diathermancy. Using ANN-FI is better than using ANN or FI alone in precision and ability of noise resistance of soft sensing. Comparisons indicate that ANN-FI is feasible in soft sensing mode ling of this kind of complex industrial process.