 This study uses feature selection scenarios and machine learning tools to develop a general model for estimating biomass higher heating value, HHV, with the highest accuracy, which is determined to be a multi-layer perceptron neural network with an absolute average relative error of 2.75% and 3.12% and regression coefficients of 0.9500 and 0.948 in the learning and testing stages, outperforming other intelligent models including a recurrent neural network recently developed using the same data bank. This article was authored by Sader Miriza Abdelahi, Saeed Faramaz-Rajbar, and Dalsa Rasegi-Joromi.