Jacobian Algorithm

[samlee2006]

I am trying to write a program in matlab to find Jaobian for any test function. As an example I am trying on Rosenbrock function and ,I think there is a minor error in the way I am defining the jacobian.
I highly appreciate if you would have a look at my Jacobian algorithm and let me know how I can change it so it will work.
-------------------------------------------------------------
rosenbrock function for 2n variables.

%rosenbrock function for 2n variables.

%function [fx] = rosenbrock(x)
function [Fx] = jrosenbrock(x)
x=[-1.2,1];

n = length(x);
fx =0 ;
Fx=[];
for i =2:2:n

f_x(i-1)=10*(x(i) - x(i-1)^2);
f_x(i)= (1-x(i-1));
%afx = 100*(x(i) - x(i-1)^2)^2 + (1-x(i-1))^2;
fx= (f_x(i-1))^2+(f_x(i))^2;


Fx=f_x';
Fx
end

-----------------------------------------------------------------------------------------------------------------------JACOBIAN CODE
epsi= sqrt(macepsi);
%fx=@rosenbrock;
Fx=@rosenbrock;
x=[-1.2 1];
%feval(fx,x);
%Fofx = feval(Fx,x);
n = length(Fx)
Jx =[];
for i = 1:n
delta = x
delta(i) = x(i)+ epsi;

for j=1:n;
%feval(Fx,delta);
%Fdelta= Fx;

Jx(j,i)= (feval(Fx,delta)- feval(Fx,x))/epsi;
end
end