Your comment suggests that you do want the results to be sorted, but according to position rather than strength. (Strength means the number of votes the circle gets in the Hough accumulator.)
There is no natural position ordering. You have to decide what you want and implement a sort as required. For example, if you want to sort by y-coordinate, you'd do this:
% example data
A = imread('coins.png');
[centres, radii] = imfindcircles(A, [15 30]);
% sort by y coordinate
[~, index] = sort(centres(:,2)); % 1 instead of 2 would sort by x coord
centres = centres(index, :);
radii = radii(index, :);
% display the ordering
imshow(A);
viscircles(centres, radii);
for c = 1:length(radii)
text(centres(c,1), centres(c,2), sprintf('%2u', c));
end
If your circles are on a kind of grid, it is unlikely that sorting by x or y will give exactly what you want. You probably have to work out which cell of the grid each circle is in, then sort the results accordingly. Here's a simple approach, but you'll need to think through how to make the quantisation step work reliably for your data.
% sort by quantised x and y coordinates
gridstep = 50; % needs to be equal to the grid size
qcentres = gridstep * round(centres/gridstep); % quantised coords
[~, index] = sort(qcentres(:,1)*size(A,1) + qcentres(:,2)); % sort on x then on y
centres = centres(index, :);
radii = radii(index, :);
You can visualise the ordering using the same code as before.
