The second-generation terrestrial transmission system for digital television broadcasting (DVB-T2) exploits Space Frequency Block Codes (SFBC) by defining a modified Alamouti code in the frequency domain in its optional Multi Input Single Output (MISO) processing transmission. Linear decoding methods such as Zero-Forcing (ZF) and Minimum Mean Square Error (MMSE) are recommended by the DVB-T2 implementation guidelines. This paper presents two modified K-Best Sphere Decoder methods to enhance the performance of the system along with low extra added complexity. The first method adapts the K-paths of the decoder according to a Channel Quality Estimation (CQE) using the Channel State In formation (CSI) of the channel while the second method utilizes the estimated Signal-to-Noise Ratio (SNR) in its adaptation process. Both methods outperform the linear ZF decoding method defined by the DVB-T2 standard in the case of selective channels.