Improper treatment of organic waste may result in environmental pollution and harm to plant growth due to the high concentration of phytotoxin. Composting has widely been used for recycling organic waste and reducing phytotoxin to improve soil properties. Assessing the phytotoxicity grades of compost products is essential for achieving high quality compost to guarantee its marketability. In this study, a technique combined parallel factor (PARAFAC) of excitation-emission matrices (EEMs) spectroscopy with projection pursuit regression (PPR) was applied to improve the sensitivity of phytotoxicity assessment during composting. Results showed that composting reduced the concentration of phytotoxin, and increased the germination index (GI). Composts were divided into four grades based on GI values. Five components containing simple DOM (component C1 and C2) and complex DOM (component C3-C5) were successfully developed by PARAFAC. Correlation analysis between phytotoxicity, chemical indices and fluorescence components demonstrated that C1, C4, complex DOM, and the ratio of simple/complex DOM components were more suitable to assess phytotoxicity of composting products. These results revealed that PARAFAC/PPR enabled a rapid and accurate method to assess the phytotoxicity of compost materials for composting plant. (C) 2016 Elsevier B.V. All rights reserved.