Coal quality information such as ash content, density, volatile matter and specific energy is important to the coal mining industry for mine planning, design, extraction, processing requirements, beneficiation and, importantly, coal type utilisation and economic value. These parameters are traditionally obtained through laboratory analyses conducted on drill-core samples collected from exploration drill This laboratory process is expensive and time consuming. In this paper, we propose a multi-variable data analysis algorithm based on the Radial Basis Function (RBF) neural network methods to estimate coal quality parameters from routinely-acquired, multiple geophysical logs, such as density, gamma ray and sonic. The performance of this RBF-based approach was demonstrated using both self-controlled training data sets and an independent data set from two mine sites. It was observed that although the density logs play a key role in coal parameter estimation, the use of multiple types of geophysical logs, including logs with different resolutions such as short spaced density log DENS and long spaced density log DENL, improves the estimation accuracy. (C) 2016 Elsevier B.V. All rights reserved.