Additive models are a natural generalization of linear parametric models, and at the same time they are able to eschew the so-called ‘curse of dimensionality’ that is often involved in high-dimensional nonparametric estimation. Traditional identification conditions and backfitting estimation technique are discussed, while these could be altered to much convenient conditions and procedures when series estimation methods are adopted. Such an advantage is demonstrated when we estimate additive nonparametric models by series methods. These models include several different types of regressors, such as deterministic, stationary, nonstationary or a mixture of them, which are commonly encountered in real data analysis for time series data. Monte Carlo simulations and empirical study are included to evaluate finite-sample properties.