A wafer processing apparatus includes a common path, extending in a Y-axis direction, in which one wafer or a plurality of wafers are conveyed, a plurality of process units stacked on both sides of the common path to constitute multi-stage structures, a main handler moved in the common path in the Y-axis direction and rotated about a Z axis at an angle theta to load/unload the wafer into/from the process units, an arm section arranged to move in the main handler in the Z-axis direction, a plurality of holding arms arranged in the arm section to constitute a multi-stage structure so as to respectively hold the wafers, each holding arm being advanced and retreated on an X-Y plane from the arm section, an optical sensor, arranged in the arm section, for detecting a holding state of the wafer in each of the plurality of holding arms, and a controller for controlling an operation of the main handler, an operation of the arm section, and operations of the plurality of holding arms on the basis of a detection result from the optical sensor, wherein the controller advances or retreats each holding arm while operating at least one of the main handler and the arm section, and causes the optical sensor to detect the holding state of the wafer by each holding arm before the holding arm reaches a corresponding one of the process units.