Identification of the signature is a process used to identify the signature of a person. The Identification of the signature can be divided into two parts. Identification signature off-line and the Identification of signatures on-line. Forgery of signatures is still common in data security systems. So, we need an approach to improve the accuracy of signature recognition on Extreme Learning Machine algorithms. This study use a gray level co-occurrence matrix (GLCM) for feature extraction and modification of Extreme Learning Machine (ELM) for recognition. ELM is a new learning method of a neural network or commonly called the Single Hidden Layer Feed forward Neural Networks (SLFNs). From the experiments the identification of the signature using the gray level co-occurrence matrix (GLCM) and signature classification using extreme learning machine with the addition of elementary transformations showed that the accuracy on identification of the signature is 43% using of features contrast, correlation, energy and homogeneity, while using just ELM accuracy is 36%.