An image forming system includes a target-log-image extracting unit and a relevant-log-image extracting unit. The target-log-image extracting unit is configured to extract a log image as a target log image likely to have been generated by use for a specific purpose of an image forming apparatus when text information extracted from the log image of the image forming apparatus by optical character recognition includes a specific phrase. The relevant-log-image extracting unit is configured to extract a log image similar to the target log image as a relevant log image based on a specific feature of the target log image extracted by the target-log-image extracting unit.