That absorbance curve is normally quite linear within its working range. If you are getting deviations from linearity then you are probably working outside of the linear range of Beer's Law. Try working with more dilute standards. Typically you won't be linear past an absorbance of around 0.8 to 0.9, depending on your spectrometer.
You will normally see the curve tail off at higher concentrations. You can typically fit a quadratic to these types of curves quite efficiently. Remember to force zero if you zeroed the spectrometer with a reagent blank (which you need to do). Quadratic curve fitting is readily available; you can find plugins for Excel on the Internet.