I have two independent variables which is affected to dependent variable directly. and there is only one moderator variable to giving effect to these two independent variable.
If the chief research question of interest has to do with whether there's any evidence for moderation in your data set, then try evaluating a regression model such as:
Estimated_Y = b0 + b1 * X1 + b2 * X2 + b3 * M + b4 * X1 * M + b5 * X2 * M
(where M is your proposed moderator).
The results for coefficients b4 and b5, if either or both were significant, would support a claim for (at least some degree of) moderation.