Processing of remotely sensed point clouds is a crucial issue for many applications, including robots operating autonomously in real world environments, etc. The pre-processing is almost an important step which includes operations such as remove of systematic errors, filtering, feature detection and extraction. In this paper we introduce a new preprocessing strategy with the combination of fuzzy information measure and wavelets that allows precise feature extraction, denoising and additionally effective compression of the point cloud. The suitable setting of the applied wavelet and parameters have a great impact on the result and depends on the aim of preprocessing that usually is a challenge for the users. In order to address this problem the fuzzy information measure has been proposed that supports the adaptive setting of the parameters. Additionally a fuzzy performance measure has been introduced, that can reduce the complexity of the procedure. Simulation results validate the efficiency and applicability of this method.