This paper explored continuous user intention through Structural Equation Modeling to find ways social robots can be used in daily life and for long-term use, such as at home. The Structural Equation Model consisted of memory and perceived likeability as latent variables, attachment as an independent variable, and continuous intention to use as a dependent variable. The scenario for the survey was created based on the description of the interviewee's attachment objects. The results of Structural Equation Modeling analysis showed that memory influenced attachment and attachment affected continuous use intention. It was confirmed that memory is an important element in the interaction between social robots and user. It suggests that experience-based services should be developed by creating long-term memories with a robot to encourage users to continue using them.