In this study a neural network for state of charge (SoC) estimation of a lead acid battery was developed. The network structure is based on a nonlinear autoregressive neural network. This network structure assures consistent SoC behaviour over time without oscillations. The inputs of the network are battery current, -voltage, derivative of the battery voltage and previous predicted SoC values. The output is the current SoC of the battery. To train the network a real-life dataset from a stationary battery storage system was used. The dataset was analysed, and a database of charge and discharge periods was created. The training dataset was developed by randomly combining the charge and discharge periods until a specified Ah-throughput is reached. The resulting current profile is used to cycle a lead acid cell under laboratory conditions to obtain the resulting voltage and SoC-responses for the training dataset. To obtain a reference SoC, the cell was discharged to 0% SOC and then charged to 100% SoC during each profile. In this case Ah-counting can be used as a reference. One exemplary training profile can be found in the attachment. Several profiles were created with the method described above and cross-validation was carried out to train and validate the neural network performance. The overall error of the method is less than one percent SoC in laboratory measurements. The algorithm also shows plausible results when used on the real-life data of the energy storage system. Although in this case no reference SoC is available to validate the performance. In the future, a method will be developed to validate the algorithm in realistic scenarios.