Strengthening and retrofitting of old structures which are being used, are important issues and significantly considered from some decades ago. There are many aims to strengthening, but according to the code recommendations, resist higher design loads, correct strength loss due to deterioration, correct design or construction deficiencies and increase ductility are the main aims. The most important problem is the debonding of FRP from concrete surface. Surface preparation is an ordinary method to prevent debonding recommended by ACI-440 Code. In this method substrate of the member must be clear and the weak or deteriorated layer should be removed by a suitable procedure such as water-jet or sanding. Wide studies have been done all over the world to investigate the failure mechanisms strengthened by this method and these mechanisms have been categorized into more than 30 groups. The most important problem is lack of a confident method to prevent debonding and yet, it threats the strengthened members. Grooving is an alternative method of surface preparation recently suggested. Although this method could improve the behavior of strengthened members and has controlled debonding phenomenon, this method must be optimized. In this paper the effect of different longitudinal patterns of grooves is investigated and three criteria including ultimate load capacity, ductility and failure mechanisms are compared for all groups of specimens. Also, the effect of thickness of CFRP plate is considered. To select the appropriate grooving patterns, it is necessary to review the studies on stress distributions along interface and determine the boundary of high stress zones.It means that grooves applied on critical areas called high-stressed zones and they omitted from zones in low state of stresses. This method has economical profits and can decrease the period of construction procedure, in addition it could improve the behavior of the members. So, 26 reinforced concrete beams by dimensions of 120x140x1000 millimeters were tested. The flexural reinforcement was 0.3 of balanced steel ratio and the 4-point load test was applied. The specimens were categorized by groove patterns and thickness of FRP layer. The beams were strengthened by 5 grooving patterns. First one is lack of grooves named EBR method. Second includes 100 percentages of groove length (Full Grooving), third includes 70 percentages of groove length (Discontinuous Grooving), fourth and fifth includes 50 and 30 percentages of groove length respectively concentrated on two ends (End Grooving). These specimens also strengthened with 3 different thicknesses of CFRP plate. To compare the ductility of specimens two criteria were determined. The first index is based on energy name energy ductility index and the second one is based on deflection named deflection ductility index. The results show that, the specimens strengthened by end grooving, was more suitable in strength and ductility than another groups of specimens and the fibers reached to their ultimate capacity, so ruptured. Also, specimens strengthened by end grooves with 150mm in length, increased ductility while decreased ultimate load. It means the end grooving can improve the behavior of reinforced concrete beams in flexure. Key Words Debonding, Surface Preparation, Grooving, Groove length Patterns, FRP thickness.