Prior literature on the impacts of children’s gender on the involvement of fathers has consistently shown that fathers spend more time with their children when they have at least one son than when they have only daughters. However, this literature has focused mostly on fathers in intact families. In the current study, I examine whether the influence of the sex composition of children on father’s involvement differs by family structure, and examine the explanations of gendered father’s involvement. Using data from the 1987-1988 National Survey of Families and Households, I find that the previously documented effect of sex composition holds only for fathers in intact families. Further examination shows that the presence of female adults in the household does not guarantee that fathers can freely display their son preference in their involvement whether there is any or not. By contrast, remarried fathers and more committed stepfathers tend to share similar child-gender effects on father’s involvement as biological married fathers in some degree. These findings suggest that highly committed fathers who believe they have a special role as a mentor for their sons may be more involved with their children when they have sons. Fathers’ differential parenting by child gender may be because of the common belief that sons need fathers more than daughters do.