The ignition, by means of a laser diode, of boron pyrotechnic mixture was carried out under various experimental conditions. The ignition system is composed of two lenses which focus the laser beam on the surface of the energetic material. A sapphire port-hole, juxtaposed with the pyrotechnic composition, protects the optical system. This setup was used to determine the sensitivity of B/KNO3 and B/KClO4 mixtures. Energy thresholds and ignition delays were measured for each batch test using a Langlie statistical method. Among the various parameters, which influence the ignition condition, the ratio of boron and the porosity of the pellet were predominant. The ignition thresholds decrease for the higher values of these parameters. The influence of an opacifier agent mixed with the mixture was found rather negligible. Boron mixture prepared with a KNO3 oxidant is much less sensitive than a pyrotechnic compound based on KClO4 oxidizer. Moreover, potassium nitrate is suspected to be unfavorable to the ignition process. Intrinsic parameters, linked to the experimental device, also have great influence. Among these the power density seems to be predominant. With the B/KNO3 mixture, a particular behaviour was observed because of the increase of the energy of ignition when the laser power is rising. It is possible to find an optimum value of the laser diameter, for a given power density. This experimental and parametric study has shown that, on such devices, it is possible to find the best conditions of the ignition of boron pyrotechnic mixture. In order to connect and explain the influence of these various parameters, a numerical model of the ignition process was completed. In the case of the B/KClO4 mixture, there is a good agreement between the numerical calculation and the experimental result. On the contrary, in the case of the B/KNO3 mixture, the agreement is acceptable until the power density reaches 2.5 kW.m(-2). Beyond this value, the numerical simulation gives very low values of the ignition thresholds of energy when compared to the experiment. The explanation of such a difference is attributed to the existence of BN particles in a gaseous medium which surrounds the pellet surface.