Wireless communications experienced a great expansion worldwide and the networks are very dense in urban areas. Predictive tools are essential to assess the distribution of electromagnetic fields in a fast and accurate way in order to optimize the implementation of base stations and minimize the human exposure. Wave propagation simulators are based on different models of urban environment and employ appropriate calculation methods. We can mention empirical, statistical, theoretical, site-specific models or a combination of them to generate a hybrid model. Given the complex and variable environment, the prediction of electromagnetic fields with such deterministic simulators has to be accompanied by an acceptable amount of uncertainty. Consequently the assessment of the electromagnetic waves in urban environment in terms of the propagation of uncertainties presents an important challenge. In order to integrate the statistical dimension into simulations and measurements, different numerical methods can be used. Here we propose the polynomial chaos expansion which is a good candidate for a parsimonious approach able to provide with various types of statistical information. This method is an advanced statistical technique which by using a polynomial expansion, provides a meta-model for the physical model.