The answer depends on what you used to fit the curve. There are several ways to do it in MATLAB.
If you used polyfit, see the documentation on polyval.
If you fitted your own function, use the fitted parameters and call your function with a vector spanning the range of the independent variable you did the fit on. It is easiest to define this vector with linspace (if you want a linearly-spaced vector).
If you used the Statistics or Optimization Toolboxes, see the documentation. For instance the Statistics Toolbox has the predict function.
