In this chapter basics of radar target classification technologies were introduced. A classification technology was presented, that decomposes a pattern recognition module of any modern radar system in the following components: Data acquisition part, signal preprocessing and feature extraction part, classification and subclassification part, data and information fusion part and finally object recognition or identification or typing part. For the data acquisition part an active or passive radar frontend can be used, that uses selfor friendly generated waveforms to reconstruct information from the environment with different objects or targets. The data acquisition part usually provides such backscattered radar echo signal in I- and Q-form in the baseband. For the signal preprocessing part some basic techniques were described in order to filter and normalise the sampled signal. It was mentioned that some measures must be taken into consideration in order to respect the basics of information theory. For the feature extraction part several basic techniques can be used. It was also mentioned that one of the most successful philosophies in designing modern radar systems for classification purpose is the best handling of the feature extraction. This philosophy consists of best understanding of the physical behaviour of a radar system in its environment. Based on this understanding characteristical feature must then been mathematically described depending on the given requirements. For this purpose the following basic methods were presented as central components of the feature extraction process: Short-Time-Fourier transform, cepstral analysis, wavelet transform and Fuzzy-logic. For the classification and subclassification part two main philosophies were presented. The first philosophy consists of learning processes. The second philosophy consists of knowledge based evidence. The different kind of classification and subclassification methods in the most modern radar systems can be divided into deterministical methods, stochastical methods and neural methods. The deterministical methods introduced in this section were essentially based on the handling of logical operators and knowledge based intelligence. The stochastical methods described in this section were based on finite stochastical automats. The finite stochastical automats presented in this section were based on different variants of learning Hidden Markov Models. Furthermore the neural methods presented in this section illustrate the capability of solving pattern recognition problems in modern radar systems by using different kinds of artificial neural networks. It was also shown that for specific classification or subclassification challenges in modern radar applications hybrid classifiers can also be recommended. This classifier uses depending on the situation learnable or non-learnable algorithms. The learnable algorithms can be designed using supervised or unsupervised learn concepts. For the data and information fusion part it was pointed out that different techniques and strategies can be used in order to fuse information from different sensor systems. It was also shown that the introduced data fusion techniques can also be integrated in a stand-alone sensor system in order to produce a robust classification and recognition result. For this purpose three technologies were presented in order to solve the given problems: Bayesian networks based method, Dempster-Shafer rules based fusion methods and finally classical rule based methods. For the object recognition, identification or typing part it was mentioned that in modern radar systems, recorded data as well as recorded intelligence information can additionally