Soil is one of the most critical parts of the environment that supplies human food. Characterizing the soil parameters, especially hydraulic ones, is crucial for protecting soil and water resources, and environmental health. Hydraulic conductivity (K) is a water movement-related attribute in the soil, which is a vital factor in managing irrigation, drainage, flood protection, and erosion control. Direct measurement of K in the field and laboratory is hard and time-consuming and needs special equipment. Therefore, numerous studies have been performed to indirectly predict this important parameter using easily measurable soil properties. However, due to its high variability, the straightforward modeling procedures normally did not predict K, accurately. It seems that complicated relationships between K and some basic and easily measurable soil parameters as predictors exist. Therefore, advanced modeling procedures are needed to discover these relationships. Artificial intelligence (AI) can recognize the complex relationships between soil parameters. On the other hand, artificial neural networks (ANNs), as a well-known branch of AI, present robust modeling procedures, which can be used for acceptable prediction of K using easily measurable soil parameters. In the present chapter, we collected and summarized the results of important studies that applied the ANNs approaches to predict K. Most of the studies in the literature predicted saturated and unsaturated K with good, very good, and excellent accuracies (0.66 ≤ R2 ≤ 0.997) using easily measurable and basic soil properties by applying both multilayer perceptron neural network (MLPNN) and radial basis function neural network (RBFNN) algorithms. While a few single studies predicted saturated and unsaturated K values with poor and fair (acceptable) accuracies (0.33 ≤ R2 ≤ 0.63) using the mentioned methodologies. In addition, saturated K was predicted with good to excellent accuracies (0.71 ≤ R2 ≤ 0.974) by inputting easily measurable soil properties and applying some lesser-used ANN-based algorithms, including group method of data handling (GMDH), adaptive neuro-fuzzy inference system (ANFIS), combination of group method of data handling and harmony search (GMDH-HS), extreme learning machine (ELM), multiple model integration scheme driven by ANN (MM-ANN), deep learning (DL), generalized regression neural network (GRNN), genetic algorithm neural network (GANN), and particle swarm optimization neural network (PSONN) in some single studies. A study also fairly predicted saturated K (R2 = 0.524) by easily measurable soil properties and using the cascade forward network (CFN) algorithm. Generally, the present chapter can lead scientists to investigate ANN-based algorithms and their capabilities to find scientific gaps in predicting saturated and unsaturated K.