In coal preparation, clean coal ash content is a key indicator for evaluating the quality of clean coal. Online ash analysers are employed to rapidly measure the clean ash, and the resulting measurements are monitored for productive process control. However, the ash observations often fluctuate and are disorganized due to random errors and analyser measurement errors, local variations in coal flow, and changing equipment and processes. Thus, this paper attempts to use the Shewhart control chart method to monitor clean ash in a statistical and global manner. From investigations of the coal preparation process, three types of variability can occur in ash process data: random variability, special variability and systemic variability. An effective X-Shewhart control chart can detect all abnormal variations under the condition that the control chart parameters (subgroup and control limits) are properly determined. To achieve this, an error analysis of an ash analyser and the influence of data autocorrelation were considered before structuring a clean ash control chart In phase I, a preliminary dataset was collected from actual ash observations. After an autocorrelation analysis, which excluded abnormal variations, a sampling scheme comparison and the selection of a population standard deviation for the preliminary data, the control chart parameters Were precisely estimated. In phase II, the phase I analysis procedure was broadly used to structure a (X) over bar chart with given standard values for monitoring clean ash. Therefore, an application example is presented, and the results are discussed to demonstrate the positive impact of this method on a coal preparation plant. (C) 2016 Elsevier B.V. All rights reserved.