This paper describes how the recently developed constrained least squares (CLS) filtering algorithm can be made iterative to improve the resolution gain (RG) of medical ultrasound images. We propose the use of the iterative CLS (ICLS) filter, by incorporating the recently proposed ultrasound tissue model, to account for the random fluctuations of the tissue signal within the received ultrasound radio frequency (RF) echo signal. The resulting improvement in RG is demonstrated by eight different abdomen ultrasound images where progressive improvements in both the axial and lateral directions can be observed.