Coronal mass ejections (CMEs) are most powerful eruptions in the solar system. They are driven by the energy explosively released from the coronal magnetic field and are often associated with solar flares, representing a dissipative energy release that causes a wide range of electromagnetic emission at different wavelengths, from radio waves to gamma rays. CMEs have strong impact on space weather - they can cause severe problems in the modern human technology and represent a significant factor in human space-born missions planning. Therefore, they are an important element of space weather forecast, which is based on a numerous ground-based and space-born observations, as well as a variety of modeling and empirical forecast methods. Namely, CMEs drive the most intense geomagnetic storms and largest short-term depressions in galactic cosmic ray (GCR) flux, so called Forbush decreases. Both of these are direct consequences of the near-Earth interplanetary conditions due to CME passage over the Earth. Currently, probabilistic forecast methods turned out to be the most efficient procedure for predicting the geomagnetic storm strength and Forbush decrease magnitude based on the remote solar observations. The presented statistical analysis reveales that both geomagnetic storms and Forbush decreases are stronger for faster and wider CMEs, associated with stronger flares originating closer to the center of the solar disc, especially when they are involved in a CME-CME interaction. Statistical relationships are employed in empirical statistical modeling based on the geometric distribution, which can provide forecast of the CME related geo- and GCR-effectiveness (i.e. geomagnetic storm strength and Forbush decrease magnitude). The evaluation reveales that the forecast is less reliable if it is more specific, and gives a relatively good prediction whether or not strongest storms and significant Forbush decreases will occur. The main advantage is in the early warning, based on the input parameters that are not necessarily satellite-dependent. Based on the presented research, two online forecast tools have been developed, available at Hvar Observatory web page. In addition, geomagnetic forecast model has been implemented in the "COMESEP alert system", which is the first fully automatic system for detection of CMEs and solar flares, forecasting the CME arrival as well as their potentially hazardous impact.