Turbo Product Codes (TPCs) with near Shannon capacity performance especially for high code rates allow decoding operations at a speed of several Mbps or more and has low error floor. In this paper, the characteristics of TPCs as FEC code, and their suitability for being a potential candidate in satellite communication applications are discussed. Emphasis is laid to highlight TPC's flexibility in attaining different code rates, improved coding gain especially for high code rates, benefits of different types of TPC construction and decoding modifications. The need for careful amalgamation of higher order modulation techniques with TPCs to achieve good spectral and power efficiency is discussed in detail. In addition, a deep learning based approach (neural-BP) for decoding of TPCs has been discussed. This is intended to improve the parallelism in TPC decoding and to reduce the implementation complexity. It is envisaged that the neural-BP decoding approach would lead to the design of hardware efficient decoders providing high Quality of Service (QoS) in satellite communication systems.