i want to reduce SSE that is generated in curve fitting toolbox, how to do this

I have points x,y,z and by curve fitting i have an equation, f(x,y) = p00 + p10*x + p01*y + p20*x^2 + p11*x*y + p02*y^2 + p30*x^3 + p21*x^2*y

+ p12*x*y^2 + p03*y^3 + p40*x^4 + p31*x^3*y + p22*x^2*y^2

+ p13*x*y^3 + p04*y^4 + p50*x^5 + p41*x^4*y + p32*x^3*y^2

+ p23*x^2*y^3 + p14*x*y^4 + p05*y^5.

For which the SSE is 3.925, my question is how to reduce this SSE?

all th 'p' terms with numbers are coefiicients having their own values , p00 = -0.08745

p10 = -0.01899

p01 = 0.02133

p20 = -8.545e-05

p11 = 9.849e-05

p02 = 0.0001887

p30 = 1.202e-06

p21 = -4.43e-07

p12 = -1.591e-06

p03 = -2.649e-06

p40 = -4.606e-09

p31 = -1.148e-09

p22 = 7.249e-09

p13 = 7.196e-09

p04 = 1.477e-

p50 = 6.537e-12

p41 = 4.849e-12

p32 = -1.26e-11

p23 = -8.424e-12

p14 = -1.338e-11

p05 = -2.795e-11

please help, very urgent

[Bjorn Gustavsson]

If that function is your model-function and those are the best fitting coefficients then that SSE is what you get. The question is if you should reduce the SSE. You could take a look at Akaike information criterion, AIC, or Bayesian information criterion, BIC, and then simlpy increase the number of model-terms until you find a minimum of either AIC or BIC. The next question is if a 2-D polynomial is the best fitting-function for your data? Perhaps you can find a better model-function?