Thanks for your request. There is no way to preserve HTML attributes after loading HTML document into Document object. Aspose.Words was designed to work with MS Word documents. Since in MS Word formats elements do not have any ids, there is no such field in the model.
Thanks for your request. Unfortunately, it is difficult to suggest you any workaround for this issue. The only possible way is parse your HTML in order to get id and index of each paragraph and then guess that indexes of paragraphs in the original HTML and in the output documents are the same. But I cannot guaranty such approach will work in 100% of cases.