In this paper, a new super-resolution ultrasonic imaging method is proposed. In this method, a set of reflected wave field data from an object is acquired by repeating the transmission and reception for all pairs of transducers on the array, then from these data the object-dependent adaptive focusing is performed in order to illuminate the desired points on the object at a constant level, and simultaneously, distributing nulls over the other regions of the object to suppress the signal. Image reconstruction is performed by scanning the adaptive focusing beam patterns over the object region and receiving the reflected wave field for this adaptive illumination. These procedures are carried out numerically by manipulating the acquired wave field data matrix and no actual beam scanning is required. Since the focusing is performed adaptively, the resolution of this method is much better than that of the extended two-dimensional maximum entropy method [Yokota and Sato, IEEE Trans. Acoust. Speech Signal Process. (April 1984)] and the active incoherent method [Yokota and Sato, Acoustical Imaging (Plenum, New York, 1982, Vol. 12; Yokota et al., J. Acoust. Soc. Am. (submitted)] for objects with fairly localized dominant reflecting regions. The effectiveness of the method is shown by numerical analysis and experimental results.