The pavement condition index (PCI), a numerical rating from 0 to 100, gives a good indication of the pavement condition. However, the pavement distress survey is a labor-intensive procedure which is performed quite subjectively by experienced pavement engineers. Then, a highly complicated calculation is required to determine the PCI of a road network. It is advantageous to determine the PCI from relevant pavement parameters. This study demonstrates how to develop a PCI assessment model based on pavement parameters by combining data mining technique and group method of data handling (GMDH) method. Records from provincial and county roads with asphalt surface and wide variety of pavement structure in Taiwan were employed. After conducting the find dependencies (FD) algorithm in data mining techniques, 120 dependent records were extracted from 253 raw records. For the PCI model development, 100 records were randomly selected as the training dataset. GMDH was successfully applied to develop a PCI assessment model that uses 7 critical pavement parameters and PCI as inputs and output, respectively. The R2 for the training dataset is 0.849. The rest of 20 records were utilized as the testing dataset, which has 0.851 of R2 based on the PCI assessment model. This study confirms that combining data mining technique and GMDH method has the potential to provide significant assistance in pavement condition assessment. The model proposed in this study provides a good foundation for further refinement when additional data is available.