Concrete is known as a durable construction material, which is widely used in many structures. In Iran due to internal production of cement and steel bars and existence of suitable aggregates, concrete is used in most of structures such as industrial floors, external floors, concrete structures, etc. In many cases, due to low abrasion resistance of concrete, durability and service life of industrial floors is low. Abrasion resistance of concrete is related to several parameters such as quality of materials, concrete mix design, water to cement ratio, curing condition and compressive strength. In this paper, the effect of some of the above parameters on ivcompressive strength and abrasion resistance of concrete and interaction of them are investigated based on statistical analysis. In addition, based on a neuro, prediction of compressive strength and abrasion resistance has been done based on statistical analysis results. Finally, some suggestions about the optimum numbers of hidden neurons have been presented. Finally, best selection of input parameters for optimizing learning time is discussed.