In this paper， an ozone prediction model which combines the Long Short Time Memory (LSTM) neural network and the Full Connection (FC) neural network (denoted as LSTMFC) is constructed， and the periodic variation rule of ozone concentration data is introduced to realize the highprecision prediction of ozone concentration. Taking the Pearl River Delta as an example， the results show that： (1) the root mean square error (RMSE) of 24hour prediction results of LSTMFC model that considers periodicity is 16.08μg/m³， the coefficient of determination (R²) can reach 0.82， which improves the accuracy by 32.28% compared with the model without considering periodicity. (2) Considering the periodicity， the LSTMFC model can achieve more accurate prediction results for the lowvalue part. At the same time， it can significantly improve the phenomenon of highvalue underestimation. The RMSE of prediction results that greater than the national primary and secondary ambient air quality standards decreased by 18.71% and 34.90% respectively， and R² increased by 40.42% and 134.04% respectively. The above results show that the proposed LSTMFC model possesses good expansibility and application potential in ozone prediction.