Aiming at the problems of complex indoor environment and poor indoor positioning precision

poor stability and high cost of traditional WiFi indoor wireless positioning algorithms in actual localization applications

and difference of strength of collected signals between different mobile terminals

a WiFi indoor positioning model based on dynamic search fireworks algorithm（dynFWA）-support vector machine（SVM） is proposed.In the process of localization

Gaussian filtering is used to perform the singular value operation on the signal and replace the traditional received signal strength（RSS） location fingerprints with signal strength difference（SSD） location fingerprints.dynFWA is used to optimize SVM parameters so as to establish dynFWA-SVM indoor positioning model.The experimental results show that the SSD fingerprint processed by Gaussian filtering can effectively improve the stability and reliability of positioning

reduce the impact on the positioning result due to the difference of signal intensities collected by different terminals. The efficiency of dynFWA is higher than the particle swarm optimization（PSO） algorithm and the fireworks algorithm（FWA）.The positioning error of the proposed dynFWASVM model is lower.