If there are only two categories in your moderating variable, then the most straightforward approach is to use that variable to form interaction effects with the continuous variables (moderation is just another name for interaction).

If you have multiple categories in your moderator, then it may be preferable to use Structural Equation Modeling (SEM).

SEM seems a prefect answer to your query.

Since your moderator is a categorical variable (for example, two groups, male vs. female), maybe you can use multigroup SEM to test and compare the mediation model for the male and that for the female group.

Thanks for reply. I have two categories in my moderating variable. Anyone of you have article related to this.

Your 1/0 variable is also known as a "dummy" variable, so look at the Sage "little green book" by Hardy, Regression with dummy variables.