Truck queuing at marine terminal gates has long been recognized as a source of emissions problems because of the many trucks that are idling. For this reason, stakeholders have a great interest in lessening the severity of the problem. To assist these stakeholders in addressing the congestion problem, baseline data and predictive models are needed. Unfortunately, data on truck queuing and research on the methodologies that can be used to estimate truck queuing time are limited. With an increasing number of marine terminals offering live webcam views of their gates to manage demand for the terminals, these webcams could be used to gather much-needed truck queuing information and other truck-related data. Data collected from the webcams were used to develop models to predict truck queuing times on an hourly or daily basis. Given the inherent fuzziness of the truck arrival data, this study evaluated the suitability of four predictive models capable of dealing with fuzzy data: multiple linear regression, fuzzy regression, clustering fuzzy regression, and support vector machines. Analysis showed that fuzzy regression outperformed other methods for the given data set.