IME, the most accurate publicly available UK weather model for the next 24 hours or so is RASP 2km:
http://rasp.stratus.org.uk/index.php/rasptable-desktop
It's crazy how good this forecast is compared to what was available 10 or 20 years ago. However, it's aimed at soaring pilots (sailplanes, hang gliders and paragliders).
Generally, if you want accuracy, it helps to compare different weather forecast models. If all models give the same forecast, you can be more confident. The more the models deviate, the greater the uncertainty.
There are many apps and websites but most free apps use the USA's GFS model (27km grid). There is no sense to compare forecasts from the same model as only the presentation and speed of updating varies.
xcweather as mentioned above uses GFS. The site also incorporates reports from airfields which allow you to compare forecast and actual.
Then there's the excellent Metoffice model which is probably the best all round choice for most people in the UK. The Metoffice app is great and worth the couple of quid for the ad free version. I occasionally look at the mountain forecasts and the synoptic charts (esp. the fax charts). The Metoffice has the advantage of being prepared / checked by human experts.
Another good one is
Meteoblue.com which is a spin off from the university of Basel and runs a custom model.
Different weather models can be compared here:
https://www.weatheronline.co.uk/cgi..._tmp2&HH=84&ARCHIV=0&ZOOM=0&PERIOD=&WMO=26038
Finally, a general understanding and observation of the weather helps. Isolated showers and thunderstorms are inherently difficult to predict for any given location. Well defined fronts vary only in the timing. High pressure can produce the same weather for days. With complex lows, almost anything could happen. Also, local effects caused by the sea or hills are rarely reflected in most forecasts and local knowledge helps here.