A spreading bush with wedge-shaped serrated leaves and large creamy-yellow flower spikes, it grows up to 3½ metres (11½ ft) high.

It has grey, fissured bark, and dark green, wedge-shaped leaves, 1+1⁄2 to 5 centimetres (1⁄2–2 in) long and 6 to 15 millimetres (1⁄8–2⁄3 in) wide, with serrated margins.

Banksia epica is similar in appearance to its close relative B. media, from which it differs in having slightly shorter leaves and larger flowers.

[3] The first European to see B. epica was probably Edward John Eyre,[4] the first explorer of the area, who recorded "stunted specimens" of Banksia as he was nearing the western edge of the Great Australian Bight on 1 May 1841: "One circumstance in our route to-day cheered me greatly, and led me shortly to expect some important and decisive change in the character and formation of the country.

It was the appearance for the first time of the Banksia, a shrub which I had never before found to the westward of Spencer's Gulf, but which I knew to abound in the vicinity of King George's Sound, and that description of country generally.

Isolated as it was amidst the scrub, and insignificant as the stunted specimens were that I first met with, they led to an inference that I could not be mistaken in, and added, in a tenfold degree, to the interest and expectation with which every mile of our route had now become invested.

[6] B. epica and B. media are the only Banksia species that occur at that location, and both have a form and habit that accords with Eyre's description.

[6][7] On 22 October, he collected a specimen of B. epica in old flower, but incorrectly identified it as B. media, and later lodged it in the herbarium at Canberra under that name.

Early in May the following year, John Falconer drove over 2000 kilometres on unsealed tracks from Warburton to Point Culver and back again, to collect fresh flowers and fruit of the purported new species.

In 1988, he published a formal description of the species, naming it Banksia epica in reference to the two "epic" journeys of Eyre and Falconer.

He considered its closest relatives to be B. praemorsa (cut-leaf banksia) and B. media, both of which have shorter flowers and smaller pollen-presenters than B. epica.

Cyrtostylis was found to be "widely polyphyletic", as six of the fourteen taxa in that series occurred singly in locations throughout Thiele and Ladiges' cladogram.

Under George's 1999 arrangement, B. epica's placement was as follows: Since 1998, Austin Mast has been publishing results of ongoing cladistic analyses of DNA sequence data for the subtribe Banksiinae.

[13] Banksia epica is known only from two populations in eastern parts of the Esperance Plains region of the South West Botanical Province, near the western edge of the Great Australian Bight.

A smaller population occurs about 70 kilometres (43 miles) further east at Toolinna Cove; when surveyed in August 1991, this locality had around 350 plants.

As Banksia species are intolerant of calcareous soils, and are not adapted to long range seed dispersal, the two populations of B. epica appear to be reproductively isolated.

The fact that the resultant isolated populations have not perceptibly speciated since then suggests that the species has been fragmented for only a short time, perhaps only since the Last Glacial Maximum.

Kevin Collins of the Banksia Farm in Albany, Western Australia is said to have pioneered its cultivation, growing it in loamy clay or sandy gravel.

[20] In the absence of further information specific to B. epica, George recommends that cultivated plants be treated as for B. media and B. praemorsa, both of which require a sunny position in well-drained soil, and tolerate only light pruning not below the green foliage.