The ability that is people’s self-monitoring and controlling of their memory processes is called metamemory. It has been regarded as truly unique characteristics of human memory, and has been studied widely as a component of metacognition in cognitive psychology. We aim to evolve artificial neural networks with neuromodulation, that have a metamemory function. Our constructive approach is based on the repetition of evolutionary experiments, analysis of the evolved networks, and refinement of the measure, to reduce the gap between the functional properties of behavior and human subjective reports of phenomenal experience. In this paper, we show the evolution of neural networks that have metamemory function based on the self-reference of memory, and analysis of the evolved mechanism of metamemory. We also discuss a similarity between the structure of the evolved neural network and the metamemory model defined by Nelson and Narens.