In this paper, we develop a spatial prisoner’s dilemma model by considering the memory use of players and apply it to a network system to evolve network dynamics using a SPD model. In our model, players sometimes ignore the classical strategy update rule and adopt a strategy that is unlikely to occur. Furthermore, players sometimes connect to or disconnect from another player if their situation is not favorable. Thanks to these effects, our model presents dynamic network evolutions with a mixture of cooperative and defective populations.