I prefer using Process Macro for one structural relationship like your case, SEM usually used in more complicated analysis with multiple structural relationships

One relevant question is whether you can assume that your measures are free from measurement error (which most measures are not).

SEM allows you to explicitly model measurement error, which usually increases power. This can be particularly helpful with moderation analysis, because moderator effects tend to be small.

Also, if at least some of your measures are based on a factor model, SEM would allow you to model this appropriately.

Dear Kogilavani Rajendran,

If you have only three variables the Process is better because SEM works with latent constructs or variables.

Also, the process is good when a moderator variable is continuous and not discrete.

Hope thgis help you

Dirceu