See the question and my original answer on StackOverflow

Html Agility Pack has XPATH support, so you can do something like this:

foreach (HtmlNode node in doc.DocumentNode.SelectNodes("//span[@class='" + ClassToGet + "']"))
{
    string value = node.InnerText;
    // etc...
}

This means: get all SPAN elements from the top of the document (first /), recursively (second /) that have a given CLASS attribute. Then for each element, get the inner text.