Japanese can be an intimidating language, but to avoid the habit of sounding like an idiot by pronouncing Japanese words like they were English words, there are some simple things to remember:
First, Japanese doesn’t usually take stress like English, so rather than emphasising one syllable, give each equal emphasis. This isn't always true in real speech but as a fundamental rule for novices this is useful.
Also, vowels. Japanese is like English in that it has five vowels, but unlike it in that pronunciation of vowels is completely uniform:
a is like in 'cat'
i is like in 'in'
u is like in 'flute'
e is like in 'egg'
o is like in 'on'
...every time. Five letters, five sounds only.
They don’t blend, so if you see two vowels together like in “Maon” or “Mie”, that’s two separate sounds, “Ma-on” and “Mee-eh”.
There are also almost no silent letters, so if you see it, it makes a sound. The only common exception is the word 'desu' (equivalent to 'is' and 'are'), which always sounds much more like "dess", with the u sound dropped. If you've seen anime in Japanese, this will be familiar even if you don't understand a word of Japanese.
The way Japanese writing works is very helpful for understanding the construction of words and how Japanese people think of them and say them. The basic unit of the language isn't individual sounds as we're used to with English letters, but syllables. Letters represent a single sound, but there are no letters in Japanese, only characters. Therefore, there's no alphabet (where character = letter = sound) in Japanese, but instead Hiragana and Katakana represent two equivalent syllabaries (where character = syllable) while Kanji are pictograms (character = idea and, flexibly, associated sound or sounds). Each character represents a syllable which consists usually of two sounds, and sometimes just one (kanji doesn't have this rule, however).
Therefore, everything in Japanese can be seen as syllables consisting of a vowel sound with a preceding consonant sound, or just a vowel sound. Syllables do not have sounds after the vowel. The one exception is 'n', which is the only vowel-free syllable. That's why the Japanese accent is the way it is and how English and other language's words get so messed around and become almost unrecognisable when spoken by heavily accented Japanese speakers. English learners born in the Japanese speaking world often really struggle to understand the idea of individual consonant sounds at the conceptual level.
L and R are the same thing in Japanese, or rather the actual sound is halfway between them, and varies a lot, so for some people it's always much more like R and for others it's always nearer L. They don't usually hear them as different either, which accounts for a lot of the spelling errors we associate with Japanese people writing English. B and V, and 'See' and 'She' are similarly easily confused due to being essentially the same sound for Japanese speakers. Certain sound combinations, particularly 'wo' as in would, wool, wolf, do not exist at all in Japanese and are therefore very hard to say, much like how English speakers struggle with some Spanish sounds like 'll', 'c' or 'z'.
A macron over a letter (a horizontal bar) indicates a long vowel; essentially a sound twice as long as normal. Since this is hard to type, this is often replaced by more familiar spelling constructions for other language speakers, e.g. Daio with a macron on the 'o' could also be written as Daioh or Daiou, it's the same word, and sometimes you see a circumflex ( a ^ symbol) used instead. There's therefore often no official spelling for a word when typing in Roman letters.
Double consonants (e.g. Yokkaichi) indicate a glottal stop. Occasionally you'll see apostrophes used instead.
Then you have the transliteration problem. There are three systems for rendering Japanese words in the Latin alphabet, of which two are very similar and one significantly different. The Hepburn system is the one that most English speakers use when learning Japanese, because it's focused on using letters in ways familiar to English speakers. This takes 土浦市役所 (つちうらしやくしょ) and renders it as 'Tsuchiura Shiyakusho' - words using three-letter syllables like shi, sho, tsu, chi, which will be read by English speakers in a way that sounds at least kind of like Japanese; I favour this system. However, Kunrenshiki is the system officially used by the government, and it works a lot more like pinyin (the system that non-native Chinese learners use to transliterate Chinese characters to phonetically-readable characters), in that it uses Latin alphabet characters in ways that are sometimes completely different from the sounds normally associated with them; letters as a code for a sound, rather than directly as sounds. Those same words, 土浦市役所 , would be written as 'Tutiura Siyakusyo'. I don't use this system, because it's stupid; to be more specific, it fundamentally fails to understand what letters are actually supposed to do. Pinyin makes sense because Chinese has no alphabet or syllabary, it's all just pictograms, and pinyin is mutually understood to be different from how English uses the letters so the weird ones are only weird for beginners. But when it's possible to use marks that map consistently to sounds, there is no excuse not to represent those sounds accurately.