Many studies on the fabrication of alternative brake pad using the natural fibers for replacing asbestos filled conventional brake pad owing to the negative impact of asbestos on the human health. Pineapple fruit is widely produced in Indonesia. Thus it can generate a lot of pineapple fiber. However, the fiber has not been utilized optimally. Pineapple fiber combined with brass powder waste and MgO powder have been potentially develop as filler of the brake material to replace asbestos. This research aims to fabricate the epoxy resin composite with three kinds of fillers (pineapple fiber, brass and MgO powders), and to study the effect of fillers ratio on the mechanical properties (tensile, hardness and wear rate) of the brake composites. The epoxy resin composites with 40% volume fraction of fillers were fabricated using a cold press method. Fillers (pineapple fiber 2 mm length, brass and MgO powders) were varied in three ratios: i.e. 50/30/20 (%), 60/20/20 (%) and 70/10/20 (%). Tensile (ASTM D 638-02 TYPE IV), hardness (ASTM E10) and wear rate (Ogoshi method) tests were conducted on all composite specimens. A scanning electron microscope (SEM TESCAN VEGA3 LMU) and an optical microscope (OM, Olympus SZ61TR) were used to characterize the tensile fracture surface and the cross-section view microstructure of the composites, respectively. The results showed that increase in volume fraction of pineapple fiber increases the wear rate, hardness and modulus elasticity of the composites. The highest value of those mechanical properties reached by the composite with filler ratio 50%:30%:20%: i.e. wear resistance 3,728x10-8 mm2/kg which is still relatively different with that of commercial break pad (8,7x10-9 mm2/kg), hardness 20,41 BHN and modulus elasticity 407,34 MPa, respectively. Besides, relatively uniform distribution of the fillers in the epoxy matrix was confirmed from the microstructure characterization. However, de-bonding and voids were still identified, leading to reduction of mechanical strength of the composites.