Ultrasound images introduce an interference pattern from noise which is known as speckle. It is speckle that diminishes its quality and leads to difficulties in diagnosis. This paper investigates fractional order differentiation and its applications in ultrasound speckle reduction. In order to preserve more texture information of image, the image de-noising model based on fractional partial differential equations was constructed by the effective combination of fractional calculus theory and partial differential equations method, and the numerical of it was achieved using fractional differential mask operator. Ultrasound phantom testing and in vivo data shows that the proposed method, compared with the traditional image denoising models, better improved the quality of an ultrasound image in the aspect of tissue SNR and CNR values.