Supraventricular tachycardia (SVT) often has very high heart rates in children. Automated external defibrillators (AED), whose use for children is recommended since 2003, could misinterpret high rate SVT as ventricular tachycardia (VT) producing an incorrect shock diagnosis. The objective of this study is to develop an algorithm to accurately discriminate SVT from VT in children. The algorithm was designed using a database of surface ECG, 322 SVT and 70 VT, collected from 259 patients in five Spanish hospitals. The mean age of the patients was 7.4 years and the mean duration of the ECG samples is 14.1 s. For each 3.2 s window two spectral parameters were calculated: the percent power content around the dominant frequency and the percent power content above 12.5 Hz. These two parameters were use to build a logistic regression model. The discrimination method was validated using a 10 fold cross validation scheme: 96.6% for SVT windows and 98.8% of the VT windows where correctly identified. The diagnosis of each window was used to compute a diagnosis for each ECG sample using a majority criterion: 96.0% of the SVT samples and 100% of the VT samples were correctly classified. Heart rate oriented AED shock advice algorithms are prone to classify high rate pediatric SVT as VT. We have developed an algorithm based on two spectral parameters that effectively discriminates SVT from VT in pediatric patients. This algorithm could be used in heart rate oriented AED shock advice algorithms to accurately diagnose high rate rhythms.