Abstarct: Like all mining activities, the blasting projects have some undesirable environmental by-products which should be controlled. These undesirable products are mainly: the blast induced vibration (whether ground or air), fly rock, produced dust and fumes. In Shahrood Cement Company, to produce limestone as a raw material for cement production, the mining activities are being preformed using drilling and blasting. The blasting projects result in producing undesirable by-product in the limestone mine and surrounding areas. One of such by products is the uncontrolled air and ground vibration which effect harmfully on the environment. Therefore the necessity of a research to record and study the vibrations seemed to be critical. The filed observations were done in the study area and all needed parameters were measured using seismograph. Following, three datasets obtained from literature were employed to extend and generalize the measured databaxse. To establish a reliable relationship between the effective factors on vibration and the produced vibration, a vibration index should be first introduced and the most effective parameters on vibration should be selected. Considering previous researches, the peak particle velocity was chosen as ground vibration index and the measured air pressure in db as air vibration index. The maximum charge per delay and distance from blast site were selected as most effective parameters on vibration. To estimate the air and ground vibration baxsed on theses parameters three approaches were used including: empirical relations, neural network and neuro-fuzzy network. Using the empirical method the best fitted equation was achieved and some standard diagrams were obtained using the suitable standards. The allowable amount of maximum charge per delay can be predicted using such standard diagrams. As an alternative, a neural network was employed to predict the peak particle velocity baxsed on the maximum charge per delay and distance from blast site. Next, a neuro-fuzzy network was used to predict the peak particle velocity baxsed on the maximum charge per delay and distance from blast site. The results of neural and neuro-fuzzy systems were satisfactory and reliable