Human gait has been commonly used for the diagnosis and evaluation of medical conditions and for monitoring the progress during treatment and rehabilitation. The use of wearable sensors that capture pressure or motion has yielded techniques that analyze the gait data to aid recovery, identify activity performed, or identify individuals. Deep learning, usually employing classification, has been successfully utilized in a variety of applications such as computer vision, biomedical imaging analysis, and natural language processing. We expand the application of deep learning to value prediction of time-series of spatiotemporal gait data. Moreover, we explore several deep learning architectures (Recurrent Neural Networks (RNN) and RNN combined with Convolutional Neural Networks (CNN)) to make short- and long-distance predictions using two different experimental setups. Our results show that short-distance prediction has an RMSE as low as 0.060675, and long-distance prediction RMSE as low as 0.106365. Additionally, the results show that the proposed deep learning models are capable of predicting the entire trial when trained and validated using the trials from the same participant. The proposed, customized models, used with value prediction open possibilities for additional applications, such as fall prediction, in-home progress monitoring, aiding of exoskeleton movement, and authentication.