Dependency parsing analyzes the syntactic structure of sentences by determining how each word depends on or is modified by other words through grammatical relationships. It parses sentences on a word-by-word basis rather than analyzing constituent phrases. Tools like WEKA can be used to generate feature files and build decision tree models to predict syntactic dependencies between words.