In order to investigate the influence of different experimental conditions and additives on the performance of ultra-light aerated concrete (ULAC), more than 300 ULAC specimens with a dry density of around 160 kg/m3 were prepared in this study. The optimum conditions of the experimental process and the dosage range of three admixtures were obtained by two series of single-factor experiments. The results show that the unification of experimental conditions is necessary for the experimental process; the range of water-binder ratio (w/b) is 0.5 ∼ 0.52; the replacement ratio range of fly ash and silica fume are (Rf, 3∼6 %; Rsi, 3∼4 %); the amount of polypropylene fiber (Mpp) is 3∼5 %. Through orthogonal experimental design, the dependence of the compressive strength, flexural strength, dry density, thermal conductivity upon Rf, Rsi, w/b, Mpp, and conditioning agent (Ac, an inorganic material) of ULAC is here investigated. The impact of four factors and their interactions on all property indexes (PIs) was studied by analysis of variance. A modified multiple-index assessment method was proposed to optimize the parameter combination for four PIs of ULAC. The results show that the Rf is the significant influencing factor of all PIs, and the interactions among the three active additives significantly impact the performance of ULAC. Nevertheless, the interaction between Rf and w/b has little impact. The repeated tests show that the deviations between the property index and the target value obtained by the theoretical prediction model are 2.63 %, 2.59 %, 2 % and 4.79 %, respectively, and according to the comprehensive weighted scoring method, the deviations between them are 2.56 %, 1.44 %, 4 % and 4.6 %, respectively. The deviations obtained by the two methods are less than 5 %, demonstrating that this method could be widely used in the ULAC multi-index system parameter optimization.