y(1)=1;
y1(1)=2;
for i=1:N
At the start of the loop (during iteration 1) y and y1 each have 1 element.
t=(i-1)*h;
k1=f(t,y(i),y1(i));
At iteration 1 this line uses element 1 of the y and y1 variables.
At iteration 2 this line uses element 2 of the y and y1 variables.
k2=f(t+h/2,y(i)+k1*h/2,y1(i)+k1*h/2);
k3=f(t+h/2,y(i)+k2*h/2,y1(i)+k2*h/2);
k4=f(t+h,y(i)+k3*h,y1(i)+k3*h);
y(i+1)=y(i)+(1/6)*(k1+2*k2+2*k3+k4)*h;
At iteration 1 this line assigns to element 2 of y, so at the start of iteration 2 y will have 2 elements.
Nowhere in this loop do you assign to any element of y1, so it will always remain 1 element long. This is the cause of the error at iteration 2. It doesn't make sense to ask for the value of the second element of a vector with one element.
end
If your goal is to solve a system of ODEs, use one of the ones included in MATLAB rather than writing your own.
If your goal is to implement your own ODE solver (because it's a homework assignment, for example) you need to determine what values to assign to subsequent elements in y1 inside the loop like you're assigning to y.
