A machine learning apparatus includes: a state observation unit that observes a characteristic shape, an adopted plan, and a determination result as state variables, the characteristic shape representing a shape of a part of a product of wire electric discharge machining, adjustment of machining conditions being deemed as necessary for the part of the product, the adopted plan being an adjustment method selected from among one or more adjustment methods for adjusting the machining conditions to improve machining performance for the part indicated by the characteristic shape, the determination result indicating whether implementation of the adopted plan is effective in improving machining performance for the part corresponding to the characteristic shape; and a learning unit that learns the machining condition adjustment method according to a data set created based on the state variables.