In this work, pillar candles are examined with eight varieties of candle wick having lengths from 1.0 mm to 19.5 mm and diameters from 1.4 mm to 3.2 mm using paraffin wax, beeswax, and soy wax as fuels. Experiments were undertaken to measure burning rate, flame height, flame width, and melt pool diameter using video and gravimetric analysis. Several properties of candle waxes were examined to understand their impact on burning rates and flame shapes. A model is proposed using the Spalding B-number with a novel wick efficiency model for burning rate accurate to ± 1.7 g/h (± 46%). The selection of wick was observed to have a strong impact on the burning rate of the candle flame. The Roper model for diffusion flames was found to provide a suitable estimate for the upper limit of flame lengths, and with the modification proposed in this work it can be used to predict candle flame lengths to within ± 10.4 mm. The width of candle flames was modelled using the Froude number and wax vapour specific gravity and was found to be accurate to within ± 1.69 mm (± 38%). The wick selection had no significant impact on the flame length or width external to its impact on burning rates. A numerical model for the diameter of the wax melt pool of a candle is presented and can be used to extend manufacturer data on melt pool sizes to other waxes. A simplified model is presented to relate the size of the melt pool to the heat release rate for various waxes.