This paper presents an algorithm for simplified features extraction by a wavelet method for off-line recognition of handwritten character, thus generating a vector of characteristics of dimension twenty one. The proposal is applied to a set of 3250 handwritten symbols, which include the digits from \(0\) to \(9\) and the upper and lowercase character of English alphabet thus giving 62 classes. The 20% of classes is used for testing whereas the rest is for training. The effectiveness of our algorithm is tested by comparison against the descriptors FKI and Wavelets by means of the classifier 1-NN. The classification is measured in percentage over the features vector set for each one of the methods. Forthermore this proposal reduce the time execution.