The continuous evolution of new technologies and the requirement for more efficient and compact power sources, justify many electrochemical researches in order to develop new types of primary and secondary power sources achieving longer operational lives and more density of energy. The primary power sources are not capable of running on charges-discharges cycles and because of this have to be discarded after the ending of electrochemical reactions (end of operational life). On the other hand the secondary sources are capable of doing charges cycles and therefore have a longer life. It is very important to realize that each kind of power source should be proper to specific applications. Though, in some cases, like emergency systems, military equipments and air and space features, it is necessary to have long shelf life sources, high reliability and high density of energy. For these special applications some sources called thermal batteries fit very well and these are primary electrochemical sources, non-rechargeable and inert at the room temperature. On the development of thermal batteries there was consolidated the technology that uses lithium-alloy anodes, eutectic salt electrolytes based on LiCl-KCl and FeS2 cathodes. The main goal of this work is to study a pyrometallugical route for the synthesis of high purity FeS2 by the reaction of vaporized sulfur and Fe2O3, for further constructions of thermal batteries prototypes. The variables of the study are temperature and time of reaction, besides the volatilizing temperature of sulfur. The analysis was done by MEV and XRD within the Rietveld method. The results showed almost 90% but a high dispersion of particles sizes. It is expected to obtain less dispersion of sizes by disaggregation methods.