Many aspects of transportation infrastructure have benefited from the use of full-scale management systems in order to quantify the system’s condition and useful life. While culvert management models have been tested in the past, none have proven functional. This project will develop a rating system for rural culvert crossing. An initial list of culvert parameters corresponding to specific failure mechanisms was reduced due to practical considerations and ease of field measurement, leaving thirty three parameters as potential predictors of culvert health. Data was collected at over 400 culverts in Montana, and was analyzed using an ordered probit statistical model to arrive at the nine parameters that were statistically significant. These nine parameters are used in the probit model to calculate a condition ranking from 1 to 5, where a ranking of 5 is best and 1 is worst. Merits of the project’s results are two fold – the exploration of the specific parameters that contribute to culvert condition are very valuable and the numerical model that resulted should be applicable in a wide range of situations.