The vast majority of researchers, actuaries, and demographers use standard time series analysis techniques to project time-varying parameters of popular mortality forecasting methods such as the Lee–Carter and Li–Lee models. However, spatial dependence can be as significant as temporal autocorrelation in these time series, and the underlying panel structure of the data is often neglected. We draw on techniques from panel and spatial econometrics, including ordinary and spatial dynamic panel linear models, spatiotemporal autoregressive integrated moving average processes, and spatial eigenvector filters, to capture such dependence and improve projections. We present a methodology to estimate the parameters of these techniques from spatial multipopulation mortality series, select their optimal hyperparameters, and use them for forecasting. We propose a tailor-made robust selection framework to identify the best model–technique combinations for each country, as well as a bootstrap-based procedure to quantify projection uncertainty with accurate nominal coverage on a separate validation period and a strategy for assessing the quality of the resulting prediction intervals. We test these methods on mortality data from 22 European countries. The results show that the proposed techniques yield a clear advantage in both point and interval forecasts for several populations, and these findings are corroborated by a robust selection design and additional robustness checks. These improvements have the potential to deliver meaningful gains for life insurance, pensions, and other contexts involving longevity risk.