This paper deals with the influence of new uses of mobile phones (such as short message service, multimedia messaging service, video, etc.) on the SAR induced in different heads models. In particular, the exposure of children will be analyzed. For this reason, three heterogeneous models have been used in this study, two children head models (9 and 15 years old) and one adult head model (“visible human”). The specific anthropomorphic mannequin (SAM) homogeneous head model has been also used to compare all the results and to confirm that the SAM model always overestimates adult and child head exposure. Two handset models have been used: the first model is a triband mobile phone having a patch antenna and the second model is a generic model (known as IEEE mobile phone) having a monopole antenna inside a metal box covered by plastic. A comparison on SAR between adult and child head is given using the finite-difference time-domain method. The results of new positions are compared with respect to the results obtained with the voice position. These studies have been performed at 900, 1800, and 2100 MHz for the triband mobile phone and at 835, 1900, and 2100 MHz for the IEEE mobile phone. The comparison of the maximum SAR over 10 g between the different heads models (9 and 15 year olds and adult) shows that results were nearly similar. It was also pointed out that the value of the maximum local peak SAR in the SAM was always higher than in the adult and children models.