Since you haven't said what you actually attempted and haven't given any sample data or even revealed the domain and range of the plot image, I can only offer a simple example.
% make up some function with a similar appearance
x = linspace(0,10,100);
y = -0.01*(x-6).^2 + 0.05*abs(sin(4*x));
% smooth the thing with an averaging filter
ys = smooth(y,20);
% smooth the thing with a spline fit (uses CFT)
fm = fit(x',y','smoothingspline','smoothingparam',0.9);
plot(x,y,'k:'); hold on; grid on
plot(x,ys,'r')
plot(x,fm(x),'b')
legend('original','average','spline','location','southeast')
You'd obviously have to tailor the parameters to whatever data you're actually using.
