There are a few web resources readily available for dealing with Latin versification (go to www.google.com and type "Latin versification" into the search box), but a few extra notes might be useful here.
Instead of getting its rhythmic effect from the alternation of stressed and unstressed syllables, Latin gets it from the pattern of long and short syllables (effectively long and short vowels). And there frequently the confusion starts, at least for English readers.
For English is richer in vowel sounds than most languages and you learn at one point to classify them as 'long' or 'short' according to the following table:

Vowel	Long as in	Short as in	Vowel	Long as in	Short as in
a	bayed	bad	e	bead	bed
i	bide	bid	o	bode	bod
u	Bude	bud	oo	booed	book

but that is not what Latin versifiers mean by 'long' and 'short' at all. For them long and short refer literally to the length of time the vowel sound is held for, the quantity of the vowel - a concept that is probably best expressed with musical notation: if a long vowel is represented by a minim (), a short one corresponds to a crotchet ().
Much of the time the two concepts correspond - saying 'bayed' or 'booed' takes longer than saying 'bad' or 'bud'. But they don't always: think of reprimanding a dog "baaad boy" or the 'a' in 'father' (at least in generally received pronunciation), which is the short 'a' sound lengthened.
Confronted by a line of English verse, there is no real way of telling what the length (in the Latin sense) of the syllables is meant to be. Take the song "Fly me to the moon and let me play among the stars". Classifying the syllables according to the table above makes no metric sense at all:

but English is flexible enough to allow the quantities to be changed at will. They could all be made equal:

or made to correspond to the natural stress:

or:

let alone the complexities that Frank Sinatra could give it.
The use of musical notation for illustration emphasises the point that Latin versification can probably best be understood by assuming that the verses are not exactly sung, but chanted as in a psalm, or delivered as in an operatic recitativo: the Germans say 'Sprächgesang' - ^speech-song'.
Take the extract from Ovid's Fasti that I have translated here. It starts

Ecce anus in mediis residens annosa puellis
sacra facit Tacitae (nec tamen ipsa tacet),

and it scans as:

a hexameter followed by a pentameter, forming an 'elegiac couplet'. I have started my translation:

Sèe the hàg who's sìtting thère amòng the gìrls
sèrving the mùte goddèss (but nòt so mùte hersèlf)

simply using a stress pattern - an iambic hexameter with a little licence - for both lines because it strikes me that this is the way it would probably have been written by an English poet.
I could have matched the stress pattern of the translation to the quantity pattern of the original:

Sèe the old hàg sitting thère in the rìng àmòng all the màidèns
Sèrving the sìlent goddèss (yet hersèlf not so sìlènt)

(One trouble here is that it drives me to use 'silent' instead of 'mute', whereas 'mute' - implying involuntary - gives the connotation of 'Tacitae' better than 'silent' - implying voluntary).
As is obvious from the Fly me examples above, having matched the stress, there would in English be no problem in matching the quantities too:

In fact, as long as the number of syllables is the same, the quantities can almost certainly be made to match: so there is no point in trying.
So the question becomes: do you match the English stress pattern to the Latin quantity pattern? Or do you try and find the natural English verse form that an English poet might use to express the same meaning? With the subsidiary question, do you do this line-by-line, or for the poem as a whole?
It is obviously a matter of taste, but for me the better solution is usually to take the second answer to both questions.
(Incidentally, with the poem in question, I accept that a better solution would be alternate iambic hexameters and pentameters, but to be frank I haven't yet come up with a translation that works in that form - in particular for the last line: a totally different problem with translating Latin is that Latin is generally denser: it packs more meaning into the same number of syllables.)