Nowadays networked control systems (NCSs) are being widely implemented in many applications. There are several problems that negatively affect and compromise the design of practical NCSs. Some of these problems are (i) data rate constrains of the channel, (ii) network traffic congestion, and (iii) inaccuracies in the model of the plant. The aim of this thesis is to develop novel noise-shaping quantizers for NCSs and their design methods that alleviate the effects of these problems. These quantizers filter the quantization errors and convert continuous-valued signals into the appropriate discrete-valued ones. First, an improved metaheuristic based design method is proposed for finite-level dynamic quantizers that minimize the performance degradation caused by quantization in systems subjected to data-rate constraints. Second, in order to deal with the network traffic congestion, a switching type dynamic quantizer that sends the data only when needed is proposed. Third, for the situations in which the model of the plant is absent or is unreliable, this study introduces a type of quantizer implemented with neural networks and a time series of the plant’s inputs and outputs. The designs of these quantizers are formulated as nonlinear and nonconvex optimization problems that are solved using covariance matrix adaptation evolution strategy and differential evolution, which are stochastic optimization algorithms. The effectiveness of these quantizers and their design methods are verified by means of several numerical examples.