The basic characteristics of a class of Hermite-type interpolation spline are investigated. Meanwhile, a new adaptive algorithm via Hermite-type interpolation spline for image interpolation is proposed by analyzing the reason of blurring or zigzag on edges of image and considering the demand for the real-time image transfer in multimedia communication. The proposed method provides the less computational complexity for a single kernel function, and the edge-preserving characters can be preserved by using the flexible border in the new algorithm. Illustrations show the validities of the proposed method by comparing with the normal polynomial methods.