Harmful algal blooms have been causing significant damage worldwide, and Hong Kong is no exception. To forecast chlorophyll-a, a proxy for algal abundance, and to understand the drivers of algal bloom formation, this study utilized nonlinear time series analysis, called empirical dynamic modeling (EDM), to investigate chlorophyll-a dynamics using in situ measurements and remote sensing data. We first demonstrated the nonlinear and deterministic characteristics of chlorophyll time series of both datasets across most sites in Hong Kong waters by comparing the predictability of linear and nonlinear models, highlighting the necessity of nonlinear approaches in studying chlorophyll-a in this region. Next, we conducted causality tests of EDM to identify significant environmental factors influencing chlorophyll-a at different sites. Our results indicated that salinity is the strongest causal factor for in situ measurement data, underscoring the importance of physical processes, such as stratification, during algal bloom formation. To assess the spatial dependence of the chlorophyll-a dynamics, we evaluated their causal relationships among different monitoring sites. Our found that, in both datasets, the chlorophyll-a concentrations in the southern sites are highly interrelated, suggesting that they share similar information. This spatial dependence occurred within approximately 30 km, indicating that a few strategically deployed monitoring sites could effectively represent chlorophyll-a dynamics in the southern region, while more monitoring sites could be beneficial in the northern regions. Finally, we incorporated the identified causal factors into the model to forecast chlorophyll-a, resulting in improved predictive performance. As satellite data showed similar results and patterns to the in situ measurements, we suggest that remote sensing data provides reliable information for chlorophyll-a dynamics even in the coastal areas. This study is the first to employ EDM to investigate chlorophyll-a dynamics in Hong Kong, showcasing its great potential. The findings provide valuable scientific insights for chlorophyll-a and water quality monitoring in real-world applications.