adilsajjad: The best approach I can think of is getting a word list, of which there are a lot on the Internet, and reading the file word by word. If the word appears in the word list, it's considered an English word.
Since word lists can be quite large structures, you may want to use something better than linear search to match candidates with the list. Possibly a sorted array with binary search lookup, or a hashmap.