Coverage prediction is an essential method of planning broadcast and long-range wireless communication systems, minimizing the need for field measurements to adjust the coverage area and consequently reducing the time and cost of implementation. Coverage prediction methods use statistical simplifications in conjunction to uncertainty quantification algorithms to assist in modeling multipath effects and impulsive noise, notably hard to be statistically considered. This paper presents a new approach to represent electromagnetic wave propagation uncertainties based on chaotic dynamical systems to model the behavior of multipath effect and impulsive noise. A novel Deterministic Chaotic Propagation Model (DCPM) is presented and used to predict coverage area in a DVB-T2 broadcast system in diverse transmission environments. The results indicate that the predicted area of bad reception is smaller in comparison to the one obtained using traditional methods. Therefore, fewer signal repeaters can be used to provide good TV reception in this area, causing a significant cost reduction in the set-up of a broadcast system. Another parameter used to evaluate the quality of reception is the carrier to noise power ratio (C/N) threshold for a bit error rate (BER) of 10-7 . The DCPM indicate a C/N threshold equal to 13.7 dB in dense urban environment and 13.4 dB in rural environment using a moderately less robust modulation configuration. Employing a more robust configuration, the C/N threshold is equal to 0.9 dB in all environments. These numbers match the obtained by the traditional propagation models.