The algorithm of NMF is so simple and elegant (just three or four lines in Matlab). Yaoliang Yu said the code can be downloaded :http://www.mathworks.com/matlabcentral/linkexchange/links/1041-matlab-code-nmf( First submitted by MATLAB Central Team on 13 Jun 2005 )
There is another version of the matlab code of NMF: http://www.csie.ntu.edu.tw/~cjlin/nmf/index.html


 The code of [http://www.mathworks.com/matlabcentral/linkexchange/links/1041-matlab-code-nmf]: 
function [w,h]=nmf(v,r,verbose)
%
% Jean-Philippe Brunet
% Cancer Genomics
% The Broad Institute
% [email protected]
%
% This software and its documentation are copyright 2004 by the
% Broad Institute/Massachusetts Institute of Technology. All rights are reserved.
% This software is supplied without any warranty or guaranteed support whatsoever.
% Neither the Broad Institute nor MIT can not be responsible for its use, misuse,
% or functionality.
%
% NMF divergence update equations :
% Lee, D..D., and Seung, H.S., (2001), 'Algorithms for Non-negative Matrix
% Factorization', Adv. Neural Info. Proc. Syst. 13, 556-562.
%
% v (n,m) : N (genes) x M (samples) original matrix
%           Numerical data only.
%           Must be non negative.
%           Not all entries in a row can be 0. If so, add a small constant to the
%           matrix, eg.v+0.01*min(min(v)),and restart.
%
% r       : number of desired factors (rank of the factorization)
%
% verbose : prints iteration count and changes in connectivity matrix elements
%           unless verbose is 0
%
% Note : NMF iterations stop when connectivity matrix has not changed
%        for 10*stopconv interations. This is experimental and can be
%        adjusted.
%
% w    : N x r NMF factor
% h    : r x M NMF factor
% test for negative values in v
if min(min(v)) < 0
error('matrix entries can not be negative');
return
end
if min(sum(v,2)) == 0
error('not all entries in a row can be zero');
return
end
[n,m]=size(v);
stopconv=40;      % stopping criterion (can be adjusted)
niter = 2000;     % maximum number of iterations (can be adjusted)
cons=zeros(m,m);
consold=cons;
inc=0;
j=0;
%
% initialize random w and h
%
w=rand(n,r);
h=rand(r,m);
for i=1:niter
% divergence-reducing NMF iterations
x1=repmat(sum(w,1)',1,m);
h=h.*(w'*(v./(w*h)))./x1;
x2=repmat(sum(h,2)',n,1);
w=w.*((v./(w*h))*h')./x2;
% test convergence every 10 iterations
if(mod(i,10)==0)
j=j+1;
% adjust small values to avoid undeflow
h=max(h,eps);w=max(w,eps);
% construct connectivity matrix
[y,index]=max(h,[],1);   %find largest factor
mat1=repmat(index,m,1);  % spread index down
mat2=repmat(index',1,m); % spread index right
cons=mat1==mat2;
if(sum(sum(cons~=consold))==0) % connectivity matrix has not changed
inc=inc+1;                     %accumulate count
else
inc=0;                         % else restart count
end
if verbose                     % prints number of changing elements
fprintf('\t%d\t%d\t%d\n',i,inc,sum(sum(cons~=consold))),
end
if(inc>stopconv)
break,                % assume convergence is connectivity stops changing
end
consold=cons;
end
end