The energy slab is a ground-coupled heat exchanger consisting of the horizontal ground heat exchanger (HGHE). The HGHE allows circulation of a working fluid through heat exchange pipes, which are buried underground with a horizontal layout, inducing heat transfer between the fluid and the surrounding ground. The energy slab is installed as one component of the floor slab layers to utilize the underground structure as a hybrid heat exchanger. However, a comprehensive design tool for predicting the thermal performance of energy slabs has not been established. In this paper, a new design method for energy slabs using an artificial neural network (ANN) is proposed and compared with the engineering chart for design. Before establishing the design method, it is necessary to accurately identify key influential parameters that may affect the thermal performance of energy slabs. Therefore, the effect of key influential parameters (i.e., the thermal conductivity of ground formations and concrete slab, the flow rate of working fluid, etc.) on the thermal performance of energy slabs was investigated by numerical analysis for ideal conditions. Then, a set of engineering charts constructed through the numerical analysis was implemented by the ANN. From this approach, the thermal performance of energy slabs for a typical condition can be readily predicted without time-consuming numerical simulations. Additionally, this new method can provide the designers with a more practical design tool comparing to the engineering chart. Note that the Bayesian Regularization algorithm with 8 hidden neurons is the optimum learning algorithm, which can maximize the accuracy of prediction.