Experiments were carried out on honey bee colonies (Apis mellifera) during the winter of 2008 and spring of 2009 using colonies wintered with five levels of varroa infestation (0.0, 3.8, 9.0, 13.1 and 22.9%) in the Central Region of the kingdom of Saudi Arabia. The goal of the present work is to determine varroa treatment thresholds and colony collapse level. Response variables examined included varroa mite infestation level (%), cm2 of brood number of frames of brood, number of frames covered with adult honey bees and mites recovered on sticky boards. The data obtained reveled that no significant differences were noticed in cm2 of brood, number of frames of brood, and number of frames of bees in colonies (range of 0.0 to 9.0% varroa infestation or 0.0 – 0.9 mites/bee). Colonies with 13.1% infestation (0.13 mites/bee) exhibited a significant reduction in all measures of performance. The present results suggested that honey bee colonies can tolerate a level of 9.0% varroa infestation (0.9 mites/bee) during the winter without evidence of colony injury. It also suggested that 13.1% infestation (0.13 mites/bee) is the colony collapse or economic injury threshold in the Central Region of the Kingdom of Saudi Arabia. Also colonies with 12.3 or more mites falling on sticky boards after 48-hour period were likely to have lower populations of bees and brood the following spring.