Hydrogen production in well a mixed system, such as a sequencing batch reactor or an anaerobic continuous-flow stirred tank reactor (CSTR) system has been extensively studied. However, little research has been focused on a large scale. The optimal environmental factors for operating the hydrogen fermentation system have also been examined by observing or traditional statistics. But the factors affecting the anaerobic biohydrogen production in a large scale were complicated, and these effects could not be explained by univariate statistics. Multivariate statistics could examine 3 or more variate (factors) in an analysis. The purpose of this paper is to analyze how well these factors could explain the bio-hydrogen production rate by multivariate statistics using a pilot-scale biohydrogen production reaction system. We found that the magnitude of the variances in the hydrogen production rate (HPR) could be explained by organic loading rate, chemical oxygen demand, volatile suspended solid, soluble microbial products, and volatile fatty acid for 77% during 65 day's observation in the pilot scale. Current results also suggested that when hydraulic retention time (HRT) was changed, the parameters were also changed. Under different HRT, the magnitude of the variances in HPR could be explained 32-77%. When HRT was set at 4 h, the variances in HPR could be explained only 32% by the possible parameters observing in this pilot scale, and the relationship between HPR and these parameters was weakest. Copyright (C) 2013, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.