Analytical methods were developed for the selective determination of alkaloids of pharmacological interest. Aiming the selective determination of camptothecin (CPT) in irinotecan (CPT-11) or topotecan (TPT) based pharmaceutical formulations, two spectrofluorimetric methods and one electrophoretic method with absorciometic detection were proposed. The spectrofluorimetric method allowed the determination of CPT using the adjustment of the alkalinity of the sample solution associated with the use of synchronous scanning or 2nd order spectral derivation. Alternativelly, the photochemical derivatization (optimized by central composite design) completely eliminated interferences on the CPT signal. The optimized conditions allowed the spectrofluorimetric determination of CPT in mixtures containing up to 50 times more TPT or up to 10 times more CPT-11. The analytical response presented linear working ranges and homocedasticity. The limit of detection (LD) was in the order of 10(-10) mol L(-1) for the method based on photochemical derivatization and one order of magnitude higher for the methods without photochemical derivatization. Tests were made using TPT and CPT-11 based comercial drugs and recoveries were around 100%. Such results were comparable to those obtained with the reference method (HPLC). A study of fluorimetric measurement uncertainty indicated that the greatest contribution in the process was the preparation of solution by volume. The contribution from this source was minimized by the preparation of solutions by weight measurement which caused a major impact in reducing the expanded uncertainty. The method based on micellar electrokinetic capillary electrophoresis (MEKC) allowed the simultaneous quantification of TPT, CPT and CPT-11. To achieve final adjustment of conditions, an univariated study was made followed by a central composite design. To improve the sensitivity of detection up to 76 times, a pre-concentration in the capillary by the normal stacking mode was used along with an extended optical path cell. The choice of borate buffer (pH 8.5) containing SDS and acetonitrile implicated in robust conditions of signal and migration times. The response showed analytical linear working range and homocedasticity, and 9 repeatability of around 3.5%. The LD was in the order of 10(-7) mol L(-1) (TPT) and 10(-8) mol L(-1) (CPT-11 and CPT). Recovery tests using spiked saliva samples were made and compared with those obtained by a reference method (HPLC) to show the appropriated accuracy for the proposed method. For the selective determination of B-carbolines, solid surface room temperature phosphorimetry (SSRTP) was used. The adjustment of the selective heavy atom and the application of synchronized scanning technique and 2nd order spectral derivation, increased the degree of selectivity, because induced phosphorescence of the analytes of interest in the sample and improved spectral resolution. The adjustment of the amount of HgCl2 (0.81 mg) allowed the selective determination of harmol in the presence of up to 10 times higher amounts of harmine, harman, harmaline and norharman. The method proposed for selective SSRTP determination of harmol was compared with the results obtained by an adapted method based on MEKC, which also showed its ability to determine harmol in the presence of interferences. The absolute limit of detection in the ng order and linear behavior of the analytical response was obtained. For the analytical method developed for the determination of harman in the presence of harmine in teas, optimization studies indicated the following conditions: AgNO3 as the heavy atom salt deposited on the substrate, sample solution at pH 11 and measurements made at 322 nm of the 2nd order derivated synchronous scanning spectrum ((delta)(lambda) = 109 nm). Tests showed that, in optimized conditions, the tea matrix had no influence on the harman signal and and that harman could be determined in samples containing harmine in concentrations up to two times higher. The analytical response showed linear range and homocedasticity. The repeatability indicated results up to 9.0%. The absolute limit of detection found for harman was 3.1 ng. In samples of threes different teas, the average recovery of harman in the presence of harmine was around 100%. The uncertainty of the SSRTP measurements indicated relevance from the four sources of uncertainty (repeatability, reproducibility, analytical curve and solutions), which is mostly caused by the variations in signals produced by non-homogeneities of the solid substrate used in measurements by SSRTP.