When studying the spontaneous emergence of cooperation within the framework of evolutionary game theory, the process of strategy selection plays a crucial role in evolutionary dynamics. In most previous studies, it has been assumed that individuals randomly select learning targets from their neighbors. However, by considering the heterogeneity of individuals’ influence in real-life social networks, a preference for more realistic choices should be taken into account. Here, we propose a preferential selection mechanism based on individuals’ payoff aspiration in memory-based evolutionary game on square lattice networks. For each iteration, players in the regular lattices will prioritize selecting neighbors with the highest satisfaction rate over the most recent runs. The simulation results demonstrate that both the level of aspiration and the length of memory can elevate the level of cooperation to a certain extent. For the moderate aspiration level, when the value of betrayal temptation is less than threshold, cooperation level monotonously decreases with the increment of memory length. However, when betrayal temptation is above threshold, memory mechanism can facilitates the evolution of cooperation, while the absence of such a mechanism leads collaborators toward extinction.