This runs without error (after several changes):
V = 2435 ;
k = 1.076 ;
Kn = 0.27 ;
k=1.076;
Y=0.39;
kd=0.04;
bs=0.15;
Xf=0.49;
Df=1.04;
Dw=1.3;
L=0.04;
a=2.24;
bt=kd+bs;
Nstar=@(N)N/Kn;
Nmin=Kn/((Y*k/bt)-1);
Lstar=L*(sqrt(k*Xf/Kn*Df))*Df*Dw;
jstar=@(J)(J/sqrt(Kn*k*Xf*Df));
Nstar_min=(1/(Y*k*(bt-1)));
nX = size(X,1 );
% solutions = cell(1, nX);
solutions = zeros(2,4);
for S=1:nX
solutions(:,S)=fsolve(@(b) Qo(S)*(1+R(S))*(No(S)-b(2))-V*(a* b(1)+((X(S)*k*b(2))/(Kn+b(2)))), rand(2,1)); % b(1) = J, b(2) = N
end
eqn1=@(b) sqrt((Lstar^1.76)+5.2*(Nstar(b(2))-Nstar_min)-(Lstar^0.88))/2.6 - jstar(b(1)).^0.88;
soln=fsolve(eqn1,rand(2,1));
It is not possible to mix symbolic functions with non-symbolic expressions. I use Optimization Toolbox functions here instead of Symbolic Math Toolbox functions.
I leave it to you to interpret the results, since I have no idea what problem the code solves.
