The performances of four multiplex PCR (m-PCR) were compared to direct immunofluorescence assay (DFA) and HuH7 cell culture for the detection of viruses in 263 children admitted to hospital with an acute respiratory illness. One hundred fifty (57.6%) nasal aspirates were found DFA-positive; 188 (72.3%) were found positive by both DFA and HuH7 cell culture, and 242 (92%) were PCR-positive. The m-PCR detected 124 viruses which were not found by conventional methods: 68 rhinovirus, 17 human metapneumovirus, 15 respiratory syncytial virus (RSV), 8 parainfluenza virus (PIV), 5 coronavirus 229E, 3 OC43 and 3 NL63, 4 enterovirus, 2 influenza virus B and C virus. The m-PCR were more sensitive, had the advantages of a shorter delay in specific diagnosis, and a lower cost than DFA and culture. Using these m-PCR, the prevalence of each virus was compared between in-patient and out-patient groups of children attending the emergency unit of the hospital. Nasal aspirates from 411 (91.5%) children were found positive by the PCRs. RSV, rhinovirus, and influenza virus were the most frequent viruses detected in this population, representing 43.6%, 31.8%, and 8.8% of the virus found, respectively, followed by human metapneumovirus (4.4%), coronavirus (3.4%), parainfluenza virus (3.2%), adenovirus (2.3%), and enterovirus (2.1%). RSVs were detected more significantly in the in-patient group than in the out-patient group, and influenza viruses were detected more frequently in the out-patient group than in the in-patient group. Moreover, the use of m-PCR pointed out the frequency of rhinovirus and mixed viral detections in these patients. In conclusion, according to the requirements of speed and low cost of the methods, and to achieve the highest rate of detection of respiratory viruses, the combined use of DFA and m-PCR is today likely to be the best way to improve diagnosis of respiratory illnesses in children.