The need to split aggregated fertility data into a fine grid of ages is a challenge that is often encountered by demographers. Several methods for addressing this problem have been developed. In this technical report, we present an application of a new approach to splitting abridged fertility data, the neural network (NN) model. Although neural networks have been widely used in various fields, they have seldom been applied in demography. The algorithm presented here is very flexible and simple to use, but it requires substantial computational resources. The NN method allows us to split abridged fertility rates of any kind using a prelearned model generated with high-quality data drawn from the Human Fertility Database. The results of testing show that in most cases, the NN model returns estimates that correspond very closely to the original values. However, the model also tends to return erroneous estimates for fertility patterns that are ‘unfamiliar’ to the pre-learned model.