Localising and tracking of a moving target is an essential and fundamental task for various applications such as inventory assets and people in warehouses etc. A well-known challenge for any wireless technology is to maintain the same localisation accuracy in both line-of-sight (LOS) and non-line-of-sight (NLOS) mixed environments. Ultra Wideband (UWB) technology has proven to be a promising technology in complex environments due to its fine timing resolution, which allows for accurate ranging estimation compared to other existing technologies. In order to achieve useful widespread implementation, issues related to NLOS propagation represent an unsolved and challenging problem set. In this work, we focus on error mitigation to improve the accuracy of the tracking and elimination of outliers in ranging data sets associated with UWB tracking systems. We first model the environment to analyse the behaviour of the signal parameters in both LOS and NLOS environments. Then, we apply a two-step adaptive algorithm for the localisation based on a Fuzzy logic approach for analysing and selecting the best ranging data to be used in the computation of the location of the device being tracked. Our proposed approach of using Fuzzy-LSE (FLSE) has been tested with real data collected in a heavy NLOS environment and has demonstrated a significant gain in the localisation accuracy, this level of improvement can reach up to 17% when compared to other well know algorithms which are described in the relevant literature.