English business negotiation requires not only grammatical accuracy but also pragmatic competence, as strategies such as directness, politeness, mitigation, and hedging determine how offers and requests are received. Current pedagogical practices, which rely on role-plays and simulations, often provide only general evaluations and lack detailed feedback on pragmatic strategies. This study proposes a rule-based approach to classify pragmatic strategies in student negotiation transcripts. Rules were formulated using both pragmatic theory and empirical observation, then represented in First-Order Logic (FOL) for computational implementation. The dataset consisted of 1,200 utterances from simulated negotiations, which were manually annotated by three experts with high inter-annotator agreement (κ = 0.82). The rule-based system classified each utterance deterministically into one of the four strategies based on lexical and grammatical markers. Results show that directness dominated (51.7%), followed by politeness (30.8%), while mitigation (11.7%) and hedging (5.8%) were used less frequently. System evaluation against manual annotation yielded strong performance, with a macro F1-score of 0.81. While effective in detecting directness and politeness, the system was less optimal for mitigation and hedging due to their implicit variability. These findings suggest that students favor straightforward expressions and underuse more nuanced strategies. The study concludes that pragmatic instruction should place greater emphasis on mitigation and hedging, while the rule-based system can serve as an automatic feedback tool to support pragmatic learning in English for Business Negotiation