Fuzzy-Probabilistic Seismic Hazard Analysis (FPSHA) is performed for Chennai city, south India incorporating both the random and fuzzy uncertainties. Randomness is handled using Monte-Carlo simulation technique, and fuzziness is accounted in the hazard analysis using fuzzy logic. The magnitude of earthquakes and epicentral distances are fuzzified and used as inputs in the fuzzy inference rules. Fuzzy attenuation relationships are developed as consequence of the inference rules with the help of ground-motion models and ANFIS. The proposed FPSHA method has the advantage over the conventional hazard analysis methods in respect of its preciseness, efficiency, practicability and reliability. The ground motions compatible with the target spectrum (UHS) of 475 years return period are selected from the recorded accelerograms with appropriate scaling. The established spectrum compatible accelerograms are vital in the seismic analysis and design of infrastructure facilities, rehabilitation and strengthening of historical and critical structures.