In this paper we present a technological study of leaf-like arrow points recovered in an open-air site from the 3th millennium B.C., named El Tossal de La Munda (Vistabella del Maestrazgo, Castellon). It is located at the central east of Spain, in the southern part of Iberian range in a mountainous area about 100 km far from the coast. The site is on a small flat-topped hill of about 5000 m² and it is 820 m.a.s.l. The upper surface has been strongly eroded and the lithic material has been found directly on the bedrock or over the dissolution clays. We recovered close to 5000 lithic remains, offering an assemblage mainly produced in the 3rd millennia BC: retouched blades and flakes as main groups, leaf-like arrow points, in addition to some other retouched tools as segments (one triangle, one bifacial circle and a trapezoid), scrapers, end-scrapers, denticulate and notches, one raclette and one drill. We also found two small adzes. The cores were intensively exhausted. Just some of them show bladelet scars but most of them are small discoid and irregular shaped cores for the production of small flakes. Few of them show marks of bipolar flaking on anvil. Within the lithic assemblage there is a sample of 73 bifacial rough-outs and projectile points at different reduction stages, from initial blanks to finished and used ones. In addition, we found one refit and one conjoint. All of them are leaf-like arrow points except a fragment showing shoulders as an exception. We have identified some typical accidents in bifacial reduction: fractured rough-outs, overshot flakes, overshot negatives and another accident that we named “bending notches”, linked principally to the pressure technique (two positives and two negatives). We analysed both preforms and finished leaf-like arrow points from a technological point of view, first trying to identify the blank type and shape, and secondly trying to distinguish knapping methods and techniques. We have divided the methods in three main types: parallels (adjacent and contiguous), alternating, and independent method. We have identified the strategies according to edges and blank faces management. In our case, only simple combinations has been detected, mainly alternate (first one face, after the other side on the same edge). We have divided leaf-like production in five technical stages, and we described the identified knapping methods and strategies used at the site according to the reduction stage. In order to show this, we describe in the paper the most significant cases of the site. The identification of blanks has been possible in most of the rough-outs, verifying that the blanks used at the site were irregular and cortical flakes, chunks, fissure slabs and small cores. Despite this, leaf-like production at the site was really homogeneous, applying the same methods and also managing the blank faces similarly. Alternate strategy is completely dominant, firstly removing the ventral face, secondly the dorsal part. The main knapping method used at the site is the independent method, removing consecutively the most highlighted ridges. When the sketch is advanced and reaches a regular shape, the application of parallel method series is common. Technically we have observed two phases: first direct percussion (mainly with stone hammer); second, pressure technique. Heat modifications have been detected in 20 rough-outs and projectiles but most of them seem to be non-intentional, exhibiting cupules and cracks. Most of the finished ones do not show double shine (heat patina). According to this, we state that leaf-like production at the site was not a specialized process but it recovers to use as blanks previously discarded flakes, chunks and exhausted cores, part of them recycled before burning. The elaboration of leaf-like arrow points was embedded in the laminar production, and it played a marginal role within a lithic reduction system that is focused on blade and bladelet production. They used mainly cortical and non-cortical flakes, that seem to be by-products of blade core preparation, or exhausted cores that were reused for this purpose. Despite most of the arrow points result in crude and thick foliated shapes, we argue that in this technological context, the use of waste as blanks constrains the knapping reduction, and lead to rude shapes. Derived from this, we discuss on the leaf-like arrow point morphology and its profitability, the resistance of thick-elliptical tips and their role in the technological framework of the site and its landscape. We also discuss about the visibility of the skilled knapping in this context, when small, irregular and cortical flakes, burned chunks and exhausted cores were used as a blank to make leaf-like projectiles.