Ephedrine and its derivatives have long been recognized for their therapeutic potential in the treatment of hypotension, asthma, and obesity. However, their use is limited by significant side effects, including cardiovascular risks and toxicity. This study presents a comprehensive computational evaluation of 13 ephedrine derivatives using integrated ADMET analysis, PASS prediction, and quantum-chemical approaches. Pharmacological profiles, toxicity risks, and electronic properties were systematically analyzed to establish structure–activity relationships. Key physicochemical parameters, pharmacokinetic properties, and toxicity risks were analyzed to identify compounds with optimal drug-like characteristics. The results highlight compound 5 as the most promising candidate, demonstrating high drug-likeness (Quantitative Estimate of Drug-likeness, QED = 0.85) and favorable pharmacokinetic properties. Conversely, compounds 3, 11–13 exhibit high toxicity and require structural optimization. PASS predictions indicate diverse biological activities, including spasmolytic and analeptic effects, with compounds 5–7 showing significant potential for further development and compound 13 exhibiting antioxidant activity. DFT analysis of ephedrine, cephedrine and key derivatives (compounds 5, 7, 13) revealed an optimal HOMO-LUMO range (ΔE = 4.7–5.0 eV) for compounds with high activity and low toxicity, where derivative 5 exhibits the best balance. This study underscores the importance of rational drug design in optimizing therapeutic efficacy while minimizing adverse effects and environmental risks.