Although E. P. Odum is generally credited as being one of the most influential holistic ecologists of the 20^th century, his ideas regarding the philosophy of ecology are in fact conflicted. Paradoxically, Odum's ecological philosophy combines both holism and the opposing belief system of reductionism. On the one hand, his beliefs are ontologically and epistemologically holistic, as demonstrated by his definition that the ecosystem is regarded as a functional whole and the nature of emergence is recognized in all levels of ecosystem. However, Odem shows a clear use of reductionist methods when creating ecological models, whereby he regards an ecosystem as a physical and cybernetic system where ecological networks are simplified into energy pathways. While Odem is, for the most part, holistic from an ontological point of view, and both holistic and reductionist from an epistemological point of view, his methods are rooted firmly in reductionist beliefs. In other words, his ontological holism is in direct opposition to his methodological reductionism, and his epistemology incorporates both opposing belief systems. This contradictory ecological thinking creates flaws in his methodology. For example, the logic in his ecological models is muddled, the evolution of ecosystems is not considered, and positivism has been excluded from his ecological research methods. Such a difficult situation could be explained by substantial speculative ecological thinking and opinions that could not be verified by ecological experiences, thus qualitative approaches, such as metaphors and descriptive language, were widely used. Otherwise, a paradox exists in ecological holism whereby it explains ecosystem complexity characteristics through an essentially putative simplicity, as if all kinds of ecosystems possess common complexity, as a highly simplified and idealized complexity. It is easy to believe that complexity characteristics are universal and normative. Indeed, the view is naturally opposed to the ontological assumptions of holists. Besides, reductionists have made tremendous achievements in classical sciences, and have helped to make ecological sciences more in line with hard sciences. Of course, reductionism in ecology is far from perfect and has not made the study of ecology entirely into that of a hard science like physics. Obviously, we cannot say that the debates between ecological holists and reductionists are futile, and we also cannot say that ontology, epistemology and methodology of reductionism are wrong. It is these controversies and improvements that have brought progress and development to ecological science. However, this does not prevent Odum's ecological thought from playing an important role in consolidating the ontological base of ecology, promoting the formation of theoretical ecology and ecological engineering, harmonizing ecological holism with reductionism and establishing a research basis for the functions of ecosystem services. From a philosophical point of view, the ecosystem is both a complicated, constitutional, generated entity as well as a relational entity with events and processes. At present, the field of ecology is on the brink of undergoing a scientific revolution. Because holism and reductionism are not sufficient to fully explain ecosystem complexity, we urgently need a new paradigm. In practice, research in ecosystem services is likely to become a model of ecology integrating other subjects, which realizes Odum's ambition of creating a modern, integrated understanding of ecology. At the theoretical level, if we are to transcend the holism-reductionism debate in ecology, the best choice is to develop a theory of ecological complexity. This can be done thorough holistic philosophy, which incorporates all methods such as positivism, metaphors and descriptive language. Indeed Odem's philosophical position regarding ontology and epistemology is in accordance with this method. To favor the formation of a unified ecological theory that effectively resolves the holism-reductionism debate in ecology, we should do the following. First, define the scale of common ecosystem research without existing ontological controversy. Second, constitute an epistemology autonomously imaging biological phenomena and laws other than physical, chemical theories. Third, implement a methodological strategy combining quantitative and qualitative means to open up new territory and generate new insights based on advanced computer models.