vekipeki - that is an awesome response...handle all the attributes as properties of a class, then serialize, deserialize the object to xml - whoa...too cool.
That said, my take on the post was the more old fashioned: I have an xml file, I want to load it from disk, find the correct element, update some attributes, and write it back to disk.
For this you can use several classes from the System.Xml library as needed:
XmlDocument to load and save a physical file from and to disk;
XmlReader/Writer to read and write, if necessary.
XmlNode and XmlAttributeCollection to find a specific attribute and modify it.
There are several similar approaches. Just to get you started, have a look at the example that comes with
XmlNode.Attributes Property
For small documents, these classes are fine. For large documents and heavy xml processing, the System.Xml.XPath library is recommended, especially the XPathNavigator.
If you can handle it, however, vekipeki's solution is the way to go.