The main objective of this work is to introduce two adaptive window-based notch detection techniques for Ultra Wideband (UWB) Frequency Coded (FC) chipless RFID system. The first proposed algorithm is the Window Based Singular Value Decomposition (WB-SVD), where various frequency dependent patterns of the notch are used as a training sequence for predefined bandwidth (window) to create an orthonormal basis functions. The WB-SVD accurately and rapidly detects the notch based on the minimum distant vector criteria. In the second algorithm, Adaptive Energy Detection (AED), an adaptive threshold is estimated for each window depending on the Probability of false alarm (Pfa) considering the frequency selective channel. Simulation results show that both algorithms are immune to environmental effects such as cluttering, fading and multipath components. Moreover, the two algorithms operated efficiently regardless on the type of chipless tags (backscattering or retransmission based). A comprehensive comparison and the trade-offs between the two techniques would be illustrated at the end of this paper.