The prognosis of type II floating knee injuries was not as good as that of type I. Our purpose is to clarify the factors affecting the outcome of type II floating knee injuries. Thirty-five patients (36 limbs) with type II floating knee injury were studied with a mean follow-up of 52 months (26-96). Blake and McBryde had classified these injuries into type I for pure diaphyseal (true type) fracture and type II if the intra-articular involvements are one or more including hip, knee and ankle joints (variant type). According to this classification, we divided these patients into two groups depending on whether their knees were involved or not. Those cases with intra-articular knee involvement were classified as type IIA, while those without intra-articular knee involvement were classified as type IIB. Of the 36 cases, 21 were classified as type IIA and 15 were type IIB. The functional outcomes of these injuries were evaluated by using the criteria of Karlström and Olerud and analyzed with multivariate analysis. After multivariate analysis with logistic regression, we show the following results: first, the poor functional outcome of type II floating knee is contributed by type IIA. Second, the type IIA group has severer femoral open fracture grading (P = 0.027) and poorer functional outcome (P = 0.009) than type IIB. Third, the significant contributing factors to final outcome are the group (P = 0.013) and the fixation time after injury in femur (P = 0.015). Intra-articular knee involvement is the most important factor contributing to poor outcome of type II floating knee. The treatment of floating knee injuries with intra-articular knee involvement is still difficult. Further efforts to search better methods of treatment are required for these complex injuries in the future.