The challenge in understanding how natural ecosystems work is linked to the complexity of ecological relationships between species, functional traits and ecosystem functions. Most of the current frameworks to studying the relationships within fundamental ecology overlook ecosystem multifunctionality, i.e., the fact that multiple species of various trait assemblages simultaneously contribute to multiple functions. Moreover, ecosystem multifunctionality has been mainly studied on the terrestrial systems of much simpler dynamics and less heterogeneity compared to their marine counterparts. Understanding the complexity within ecosystem multifunctionality in the real-world complex systems is a critical step to be able to improve our knowledge and management of socio-ecological systems. In this thesis, I develop different ways of characterising ecosystem multifunctionality in coastal marine systems. The frameworks that I developed are Network Analysis of Traits, Bayesian Belief Multifunctionality Framework and Multivariate Network Analysis. These frameworks incorporate network theory that supports a holistic and transparent analysis of multifunctional relationships and provides insights into changes in community composition, changes in functional diversity, redundancy patterns, ecosystem resilience and functional synergies and trade-offs that can result from forcing an ecosystem in a particular direction. The frameworks were tested on real-world complex coastal ecosystems, namely, the intertidal sandflats of Kaipara Harbour and Whangateau Harbour, New Zealand. This thesis begins with Introduction, where I explain the background of my research and highlight the research gaps. Chapter 1 of my thesis presents Network Analysis of Traits that looks at the changes in the community structure subject to an increased level of experimental nitrogen. Chapter 2 shows Bayesian Belief Multifunctionality Framework in the assessment of multifunctional relationships between species, functional traits, ecosystem functions and nature's contributions to people. Final Chapter 3 shows Multivariate Network Analysis developed based on empirical data collected by me and shows the role of individual trait clusters and environmental characteristics in explaining multiple ecosystem functions that I measured. The frameworks that I created provide insights into the mechanistic ecological underpinnings of complex multifunctionality relationships and will support increasing our knowledge of how natural ecosystems work. Such knowledge is fundamental to the successful management of socio-ecological systems.