@Arjun I suggest you to use Process macro for SPSS to run moderation analysis because AMOS doesn't offer moderation analysis function. To run moderation in AMOS, you'll have to run path model with IV, Moderator, and Interaction Variable (product of IV and Moderator). Moreover, AMOS doesn't work with categorical variables as well. Another option for moderation analysis along with your path model would be Smart-pls and it works well for categorical variables also.
For a dichotomous nominal variable, just code the values as 0/1, or 1/2, and proceed. For a nominal variable with more than two levels, you'll need to recast the variable as c - 1 dummy variates (where c is number of levels); each can be coded 0/1. You'll need to include all of the dummy variates in order to carry the explanatory information in the nominal variable.
For example, if you recorded marital status as: single, married, divorced, widowed, then dummy 1 could be coded 1 if the person was single, and 0 otherwise; dummy 2 could be coded 1 if the person was married, and 0 otherwise; and dummy 3 could be coded 1 if the person was divorced, and 0 otherwise. Note that each category will have its own unique pattern on the three dummy variates (single = 1, 0, 0...widowed = 0, 0, 0).