The sensors reliabilities in multisensor distributed decision systems with data fusion are the a priori statistical data needed to optimize the system it terms of higher detection probability. This paper proposes a new approach in data fusion systems to estimate the sensors reliabilities and to improve the system performance. The proposed approach is based on neural network and fuzzy logic technologies. Use of neural networks to learn system behavior seems to be a good way to solve the problem of the needed a priori statistical data in multisensor distributed detection systems with data fusion. Also, fuzzy logic has been proven very successful in solving problems in many areas where conventional model is either very difficult or inefficient/costly to implement. Use of fuzzy logic in multisensor distributed detection systems to determine uncertainty or confidence value (grade of membership function) for each hypothesis has the advantages of the soft decision approach. Combining the two technologies in multisensor distributed detection systems provides the benefits of both technologies. Thus using neural network and fuzzy logic technologies reduces the needed statistical data and improves system performance. The proposed approach does not require a priori statistical knowledge of the sensing process. The optimum fusion rule using the proposed approach is derived. The performance of the proposed approach is evaluated and compared to the performance of the hard-decision approach. The proposed approach provides detection probability improvement over a comparable hard-decision system and is able to correctly estimate the sensors reliabilities.