Earthquakes have a damaging impact on the economic welfare and resilience of communities, particularly in developing countries. Seismic hazard assessment is the first step towards performing prevention, preparedness, and response or recovery actions to reduce seismic risk. This paper presents a computation tool for predicting the seismic hazard at the macro level as a part of a comprehensive multi-hazard framework on earthquake risk assessment. The probabilistic seismic hazard analysis (PSHA) procedure is based on the Monte-Carlo approach, and particular attention is paid to the definition of source zones assigned in the study area. Both Poisson and time dependent (renewal) models are adopted to quantify the effect of temporal dependencies between seismic events, while near-field rupture directivity effects are also taken into account. Marmara region in Turkey is selected as a case study area to perform a new seismic hazard analysis and verify the accuracy of the proposed tool. The results show good agreement with results from the recent SHARE project and the latest Turkish Earthquake Design code hazard maps. This confirms that the proposed PSHA method can be an attractive alternative to the direct integration based methods due to its practicality and powerful handling of uncertainties.