Precise automated determination of char morphologies formed by coal during combustion can lead to more efficient industrial control systems for coal combustion. Commonly, char particles are manually classified following the ICCP decision tree which considers four morphological features. One of these features is unfused material, and this class of material not characteristic of Colombian coals. In this paper, we propose new machine learning algorithms to classify the char particles in an image based system. Our hypothesis is that supervised classification methods can outperform the 4 ‘class’ ICCP criteria. In this paper we evaluate several morphological features and specifically assess the contribution of the unfused material feature on the overall classification performance. The results from this work confirm that the proposed method is able to accurately identify and automatically classify chars.