As a simplified approach to optimize the production of formaldehyde from partial oxidation of methane, this paper illustrates the application of statistical multivariate analysis and prediction through a multi-regression model, as well as the optimization of the processing parameters in the complex reaction process. This approach is advantageous, especially when experimental evaluation and optimization of a process is time consuming and expensive. By carrying out a finite number of experiments, statistical modeling in this work shows reasonable good prediction ability in terms of methane conversion, formaldehyde selectivity, and space time yield (STYHCHO). Moreover, not only are the synergistic effects between reaction parameters revealed, but also a comprehensive understanding of the whole production process over variation of the all reaction parameters (within the phase space represented by the experimental grid) are allowed. (c) 2006 Elsevier B.V. All rights reserved.